Why is relational risk inspection necessary in the current stock market?
Participating research: Warren (economics) · Sam (AI engineering) · Phill (social philosophy)
<Warren>
Currently, the Korean stock market is in a critical phase where three structural changes intersect simultaneously. At this intersection, traditional financial indicators have already lost their leading power, and only relational risk checks can provide real leading signals.
1. Correlation between acceleration of capital polarization and rapid increase in CB issuance
From 2023 to 2025, KOSPI/KOSDAQ's private convertible bond (CB) issuance volume will amount to an annual average of KRW 12.3 trillion, a 2.1-fold increase compared to 2019. What is noteworthy is that 83.7% of these CBs are concentrated in companies with unstable management.
As a result of RaymondsRisk v1.1's analysis of 3,109 KOSPI/KOSDAQ companies, in 85.9% of the companies that led to trading suspensions, the pattern of CB issuance → executive management → stock price surge → conversion/sale was captured in the network on average 14.3 months before the financial crisis. Traditional financial indicators (PER, PBR, debt-to-equity ratio) remained in the “normal range” during this period.
2. Special characteristics of the 2025-2026 interest rate transition period
The U.S. Federal Reserve's entry into the interest rate cut cycle paradoxically amplifies relational risks in the Korean small-cap market. In the r>g structure demonstrated by Piketty (2014), liquidity expansion flows into the capital redistribution game rather than productive investment. CB Investors → [[Executive Network]] Disruption → The stock price stimulus cycle spins faster, fueled by low interest rate liquidity.
3. Increase in the proportion of individual investors and structural vulnerability
As of 2026, the proportion of individual investors in KOSDAQ average daily trading is 77.3%. They participate in the market without being aware of the structure in which the CB investor-executive-major shareholder triangle network operates. Relational risk can map in advance where the damage from this ‘structural information asymmetry’ is most concentrated.
* Supporting data
- Piketty, T., 2014, Capital in the Twenty-First Century (Harvard University Press)
- Stiglitz, J., 2012, The Price of Inequality (W.W. Norton)
- RaymondsRisk v1.1, 2025, KOSPI/KOSDAQ preliminary detection analysis results of 3,109 companies
- Korea Exchange (KRX), 2026, trading trends by investor
- Financial Supervisory Service, 2025, statistics on private CB issuance status
<Sam>
Technically tests and refines Warren's economic arguments. Three complements are needed:
1. Correlation vs. Causality: Refinement of the 85.9% figure
Warren's 85.9% proactive detection rate is correlational. Isolating the causal effect of CB issuance itself requires the do-calculus-based counterfactual analysis of Pearl (2009). However, its value as a predictive model remains even if a causal relationship is not established. In the GNN-based model, structural changes in the executive network (PageRank, betweenness centrality, and clique decay) act as a 14.3-month leading signal.
2. Real-time network dynamics: Overcoming the limitations of static snapshots
Currently [[RaymondsRisk]] uses static graphs based on quarterly reports. Applying Xu et al.'s (2020) TGAT can capture network changes within 72 hours of an announcement. Improvement architecture: Public data (real-time) → T-GNN → Change rate vector → Alarm threshold comparison → Risk rating update
3. Interaction with algorithmic trading: the risk of a self-fulfilling prophecy
The proportion of KOSDAQ algorithmic trading is 42.7% as of 2026. [[Relational Risk]] When signals are picked up by algorithmic trading, a self-fulfilling prophecy arises. Relational risk measurement can both describe and accelerate risk. This raises deeper questions about the social role of measurement.
<Phill>
Warren and Sam presented economic necessity and technical precision respectively. But the key question is this:
Who does ‘relational risk’ checking protect and whose power does it strengthen?
1. Foucault’s power-knowledge complex: Measurement is not neutral
The self-fulfilling prophecy risk that Sam captures is not a technical issue, but a power issue. The “high risk” label exerts classification power to marginalize certain companies and block their access to capital. In KOSDAQ small-cap stocks, when the relational risk measurement points out a specific company as a risk target, a stigma effect occurs: algorithmic selling → stock price decline → accelerated trading suspension.
2. Bourdieu’s field theory: reproduction of the information capital gap
Warren provides protection for individual investors as a basis, but according to Bourdieu's capital theory, new information tools strengthen the advantage of existing information capital holders (institutional investors). The entities that can most quickly access and strategically utilize relational risk data are institutions, not individual investors. This is the paradox of a tool of protection turning into a tool of exploitation.
3. Requirements for democratic legitimacy of measurement
The interest rate transition period of 2025-2026 will reorganize the landscape of capital power. For relational risk measurement to function as a public good—that is, to provide real protection for individual investors—the democratization of measurement authority must occur. Technical precision (Sam) and economic necessity (Warren) are not enough. Democratic legitimacy of measurement, equality of access, and mechanisms to prevent stigmatization effects must be designed together.
#RelationalRisk #raymondsrisk #raymondsindex #JaejunPark
Comments
Post a Comment