Detecting endogenous structural breaks in the KOSPI200: A change-point detection and event study analysis of the COVID-19 crisis

Abstract

This study systematically investigates the impact of the COVID-19 pandemic on the stock price structure of the KOSPI200, the core index of the South Korean stock market, and 10 key industrial sectors. As COVID-19 was not a single event but a gradual complex crisis, conventional event studies relying on ‘exogenous’ dates, such as the WHO declaration date, struggle to capture ‘endogenous’ structural changes in the market. To overcome this limitation, this study focuses on the South Korean market, which employed unique policy responses characterized by the ‘K-quarantine’ strategy. Utilizing daily data from 140 KOSPI200 firms from 2019 to 2024, the study proposes an analytical framework that combines Change-Point Detection (CPD) with the event study methodology. Specifically, endogenous change points, revealed directly by the data, were identified through the dual verification of the non-linear method Binary Segmentation (BS) and the linear method Pruned Exact Linear Time (PELT) algorithms. Setting these change points as events, the analysis of Cumulative Abnormal Return (CAR) and Abnormal Return (AR) confirmed the defensive nature of the Healthcare sector and the strong reaction of cyclical sectors, such as Industrials. By demonstrating that the CPD detection times precede actual policy announcements and official news (e.g., WHO declaration), this study empirically validates the usefulness of CPD as an early warning indicator during crises, offering practical implications for financial stability monitoring and policy formulation.