Information Production by Institutions and Information Extraction by Underwriters in Hybrid IPO Auctions

The optimal mechanism for selling shares in Initial Public Offerings (IPOs) has been widely debated. On the one hand, auction theory suggests that IPO auctions are an efficient way to sell shares under a wide range of conditions (see, for example, Ausubel, 2002). On the other hand, as documented by Jagannathan, Jirnyi, and Sherman (2015), despite being adopted at various times in at least 50 countries, IPO auctions have survived in only a handful of these cases, having largely given way to fixed-price offerings or bookbuilding. Under bookbuilding (now the dominant IPO mechanism in the U.S.), the IPO underwriters rely heavily on institutions’ ability and willingness to produce information about a firm’s value before it goes public. Underwriters collect bids and qualitative feedback during roadshows and, crucially, retain discretion over allocations. That discretion allows them to reward investors who produce more precise information. But it also creates room for abuse, particularly in countries where regulatory oversight is weaker. Excessive commissions, laddering (conditioning IPO allocations on investors’ commitments to purchase shares in the aftermarket), analyst conflicts, and spinning (allocating IPO shares to executives’ personal accounts to influence their firms’ investment-banking decisions) are not theoretical possibilities; they are well-documented features of poorly governed bookbuilding regimes.

This trade-off has motivated interest in alternative IPO designs that reduce discretion without destroying information incentives. One such design is the hybrid IPO auction, in which an initial uniform price IPO auction is conducted with participation limited to institutional investors, followed by a retail tranche in which retail investors are allowed to buy shares at a fixed IPO offer price determined by the underwriters based on the information they may garner from the IPO auction. China has required this hybrid structure since 2009, making it one of the few large markets to implement an auction-based IPO system at scale.

The Chinese hybrid mechanism differs from bookbuilding in a critical way. Underwriters no longer control allocations. Instead, institutional investors who bid above the offer price receive shares on a pro rata basis. At the same time, retail investors are confined to a separate fixed-price tranche, reducing the risk that uninformed retail bidding distorts auction outcomes. In principle, this design curbs allocation-related abuses while preserving a role for institutional information in price discovery. But a key concern remains: if underwriters can no longer reward institutions that produce high-quality information, will these institutions still invest the effort? And if they don’t, might IPO prices end up drifting further from their true value?

Thus, whether hybrid IPO auctions are a suitable mechanism that can accomplish the objective of pricing IPO firm shares close to their intrinsic value is an empirical question. Notably, there has been little systematic investigation of this issue in the literature. While prior research has carefully examined information production under bookbuilding, there has been no comparable analysis of hybrid IPO auctions.

Our paper fills this gap by studying China’s hybrid IPO auctions using a uniquely rich dataset of more than 115,000 institutional bids across 715 IPOs conducted between 2009 and 2012. These data allow us to observe not only bid prices and quantities, but also allocations and detailed institutional characteristics, such as whether an institution is domestic or foreign, geographically close to the issuer, large or small, or a mutual fund versus another type of financial institution.

We empirically investigate the following research questions that go to the heart of IPO design.

First, do underwriters actually use institutional bidding information when setting offer prices in hybrid auctions? The answer is clearly yes. Offer prices are strongly related to both average bid prices and aggregate demand in the auction. More importantly, underwriters place greater weight on bids from institutions that appear to produce more precise information, such as domestic institutions, geographically proximate investors, larger institutions, and mutual funds and securities companies.

Second, does institutional bidding predict IPO initial returns? Again, the evidence is affirmative. IPO initial returns increase with both the bid-to-price ratio and oversubscription, and the predictive power is substantially stronger for bids from institutions with higher information precision.

Third, does institutional bidding contain information about long-run performance? We find that it does. Institutional bids predict abnormal returns not only in the immediate aftermarket, but also over longer horizons—up to three years after the IPO. This is consistent with institutions strategically using their information over time, especially in the presence of retail investors who may contribute to short-term mispricing.

Finally, are institutions that produce better information actually rewarded in a system without discretionary allocations? Here, the results are particularly striking. Institutions that are able to produce more precise information receive larger allocations in better-performing IPOs and earn significantly higher profits. Even without underwriter discretion, information production pays.

We also address alternative explanations, including underwriter favoritism and belief dispersion among investors. Regulatory reforms in China that increased post-IPO disclosure provide a natural experiment, and the evidence points squarely toward information production—not favoritism—as the primary driver of our results.

In summary, we believe that an important contribution made by our paper is to enhance our understanding of the appropriateness of alternative IPO mechanisms for use in various countries (making use of the Chinese IPO mechanism and economic setting as a real-world laboratory that we can learn from). Our analysis suggests that many developing countries can benefit from adopting a two-stage hybrid IPO auction as their IPO mechanism. Such a mechanism, by requiring a fixed-price retail tranche, can ensure some equity between large and small (retail) investors in terms of their access to the IPO market, while at the same time eliminating potential distortions caused by having both institutional and retail investors participating in a common auction tranche. In particular, the results of our analysis of information production by institutions suggest that the switch to a hybrid IPO auction mechanism need not entail a sacrifice in information production by institutional investors relative to that in the bookbuilding IPO mechanism now prevailing in the U.S. and many other developed (especially European) countries.

-----------------------------------

Thomas J. Chemmanur is a Professor of Finance in the Seidner Department of Finance at the Carroll School of Management of Boston College.

Pengfei Ma is an Assistant Professor of Finance at the Lee Kong Chian School of Business, Singapore Management University.

Chaopeng Wu is the Dean of the School of Management and a Professor of Finance at Xiamen University.

Qianqian Yu is an Assistant Professor of Finance in the Perella Department of Finance, College of Business, at Lehigh University.

This blog is on the paper presented at the 6th Annual RCF-ECGI Corporate Finance and Governance Conference, held in a hybrid format online and in Hoboken, New Jersey, on 13 and 14 December 2025. Visit the event page to explore more conference-related blogs.

The ECGI does not, consistent with its constitutional purpose, have a view or opinion. If you wish to respond to this article, you can submit a blog article or 'letter to the editor' by clicking here.