Standards for setting the event period and estimation period under the event study method in market manipulation damages claims

Cross-family AI independent reading & axis check

5 of 8 axes matched GPT-6 Astra (OpenAI, codex CLI) · 2026-10-16

When compared with an independent reading by another AI family (GPT-6 Astra), which read only the original text without seeing the initial reading, 5 of the 8 axes matched, and the readings diverged on Axis 6 (Relation to precedent), Axis 7 (Burden of proof), and Axis 8 (Open question and revisit condition).

Raw responses and comparison record: Protocol · Blind readings (JSONL) · Axis comparison (JSONL) · SHA-256 (JSON)

Lower-court contrast & scope of application

Lower-court reading vs. Supreme Court holding

Lower-court context: The court below did not credit the appraisal result that the stock price had fluctuated under the influence of the market manipulation and that the plaintiffs had thereby suffered loss, and, finding no other evidence establishing adequate causation, did not accept the plaintiffs' claim. The appraiser set September 1, 1996, approximately four months earlier than the commencement date of the market manipulation, January 3, 1997, as the start date of the event period, and set the period from November 20, 1995 to July 26, 1996 as the estimation period.

Supreme Court standard: The Supreme Court found no rational basis for the appraiser's setting of the start date of the event period, nor for the setting of an estimation period that included the period preceding the structural change in which the value of the company's real estate holdings came into prominence around April 1996, and held the determination of the court below to be justifiable. It also held that, given that the defendant's separate analysis results had been submitted as documentary evidence, the court below's failure to order a re-appraisal ex officio did not constitute a failure to exhaust deliberations.

Distinguishing conditions

• Cases in which special circumstances related to that act, such as the accumulation of shares in connection with the market manipulation, are established by evidence before the commencement date of the market manipulation
• Cases in which no structural change affecting expected returns occurred during the estimation period, so that the reliability of the regression equation is maintained
• Cases in which causation and loss are contested under the statutory provisions for estimating the amount of loss or by another method of calculation, rather than the event study method

Record verification notes

Before asserting the same conclusion, it needs to be checked whether the start date of the event period in the appraisal coincides with the commencement date of the market manipulation, whether the estimation period mixes in a period preceding a structural change specific to the company, and whether an industry index appropriate as an independent variable in the regression model has been selected.

From the official text

Points at issue

주식에 대한 시세조종행위와 그 주식의 매매거래나 위탁을 한 자가 입은 손해 사이의 상당인과관계 유무를 판단하기 위하여 이른바 사건연구방식의 분석을 활용하는 경우, 사건기간과 추정기간의 설정 방법

Summary of the holding

The official text publishes no summary of the holding for this decision. The quotation and factors in this record are copied verbatim from the points at issue; the evidence for each axis is taken from the points at issue, the order and the reasons.

Order

상고를 모두 기각한다. 상고비용은 원고들이 부담한다.

Provisions referred to

증권거래법(2007. 8. 3. 법률 제8635호 자본시장과 금융투자업에 관한 법률 부칙 제2조로 폐지) 제188조의4, 제188조의5 제1항, 민법 제750조, 제756조