IPO pricing efficiency has been a hot topic in academic research at home and abroad, and it is also the key content of China's stock market issuance system reform in recent decades. This paper focuses on whether IPO pricing can effectively reflect the intrinsic value of the company, and empirically tests the IPO pricing efficiency of the main board market by using the stochastic frontier model. On this basis, the mean regression and quantile model are used to test the impact of various factors on the pricing efficiency from the perspective of the mean and conditional distribution. The research finds that the average IPO pricing efficiency of the main board market before the full registration reform is 81.43%, and the average IPO pricing efficiency is 88.78% after the registration reform. After the registration reform, the pricing efficiency of listed companies on the main board market has been improved, and the IPO premium has been alleviated. In addition, the estimated results of the mean and quantile model show that the main underwriters ‘only recommend not guarantee’ and moral hazard problems are still prominent, and the sponsor-follow-investment system has no obvious constraints on the top brokerages. But the removal of the implicit restriction of 23 times the issuance price-earnings ratio has a significant improvement effect on the companies with high pricing efficiency. Therefore, the regulatory authorities should pay more attention to the IPO price-earnings ratio of IPO companies and the moral hazard of the main underwriters. At the same time, the construction of a symmetrical incentive and restraint mechanism for market players is still the core content of the improvement of the system.

Comprehensive Registration System Reform, IPO Pricing Efficiency, Stochastic Frontier Model, Quantile Regression Model