Abstract:
Since the implementation of the registration-based system, the phenomenon of excessive volatility in the early stages of IPO listings persists, affecting market pricing efficiency and rights and interests of small and medium investors. Using a sample of 1, 162 IPOs under the registration-based system, this paper analyzes the impact of individual investors' cognitive biases on IPO pricing efficiency. It innovatively designs characteristic indicators of investor cognitive bias and early warning tools for IPO low-pricing efficiency from the perspectives of the book-building process, characteristics of companies to be listed and underwriters, and post-listing trading behaviors. The findings show that: (1) From a volatility perspective, 96.3% of IPOs exhibit excess volatility that cannot be explained by market style and is solely caused by listing time differences when compared to stocks in the same industry with similar fundamentals. (2) From a return perspective, secondary market premiums constitute the main component of IPO low-pricing efficiency, with a variance contribution as high as 79.2%. (3) Investors excessively focus on price-volume characteristics in the book-building process and the magnitude of changes in offering prices relative to median inquiry prices, while ignoring important information related to the intrinsic value of IPOs. This, combined with the reinforcing effects of underwriter reputation and trading behavior of some "IPO speculation" investors, generates irrational overvaluation in IPO secondary market prices. (4) During the book-building process, issuers, underwriters, and inquiring institutional investors may collude to manipulate by window-dressing the intrinsic value of IPOs and altering the overall distribution of quotes, triggering cognitive biases such as representativeness bias, absolute price illusion, and confirmation bias among investors. (5) The IPO low-pricing efficiency early warning tool constructed based on GLS methods and XGBoost machine learning models identifies 22 indicators characterizing cognitive biases, achieving a prediction accuracy of 93.8%. This paper enriches research on IPO pricing efficiency and provides references for optimizing market resource allocation and strengthening investor protection.
下载:
