Heterogeneous stock market responses to pharmaceutical policy events across subsectors and time scales: Evidence from China.

Whether abnormal returns around different pharmaceutical policy events exhibit similar patterns, and whether these patterns vary across industry-index levels, return scales, and event windows, has received limited comparative examination within a unified analytical framework. This study examines six pharmaceutical policy events in China using the Shenwan first-level Pharmaceutical and Biotechnology Index and six Shenwan second-level pharmaceutical sub-sector indices, with the CSI 300 Index as the market benchmark. Ensemble Empirical Mode Decomposition (EEMD) is used to reconstruct returns at multiple scales, and an event-study framework is applied to estimate abnormal returns (ARs) and cumulative abnormal returns (CARs). The results show that, within the events, industry indices, and event windows examined, policy objectives or policy-instrument attributes do not exhibit a consistent correspondence with observed abnormal-return patterns. Medium-frequency CARs were generally negative for both centralized procurement events, but centralized drug procurement showed broader negative accumulation across sub-sectors, whereas centralized procurement of high-value medical consumables showed greater sectoral differentiation. Although the clinical trial data self-inspection and verification event and the reform of the drug review and approval system both involved drug R&D and registration regulation, their abnormal-return patterns differed markedly, with the latter showing broadly positive medium-frequency CARs. The aggregate industry index may not fully capture underlying differences when sub-sector responses diverge substantially, and high- and medium-frequency CARs also differ in direction for some events and sub-sectors. In addition, the cumulative direction of CARs over wider event windows may differ from the trajectory of daily ARs following the policy announcement. Overall, generalizations based on policy categories, an aggregate industry index, or a single event window may overlook important differences in abnormal returns across industry-index levels, return scales, and temporal paths.