COME ON IN, THE WATER'S FINE! AN EXPERIMENTAL EXAMINATION OF HYBRID IPO AUCTIONS WITH A PUBLIC POOL

• Published date: 01 March 2019
• Author: Minjie Shao,Ann E. Sherman,Charles R. Schnitzlein
• Date: 01 March 2019
• DOI: http://doi.org/10.1111/jfir.12169

COME ON IN, THE WATER’S FINE! AN EXPERIMENTAL EXAMINATION OF

HYBRID IPO AUCTIONS WITH A PUBLIC POOL

Charles R. Schnitzlein

University of Vermont

Minjie Shao

Tusculum University

Ann E. Sherman

DePaul University

Abstract

We offer experimental and theoretical evidence that auction methods for initial public

offerings (IPOs) in the United States may be improved through hybrids that include a

separate retail tranche or “public pool.”The multi-unit uniform price auctions in our

laboratory experimentsincorporate key IPO features includingendogenous entry, costly

informationacquisition, and differing capacityconstraints. As predicted,hybrids increase

proceedsand reduce price volatility, price error, and the incentivefor small bidders to free

ride. Underpricing occurs in bothtreatments but is attenuated with the public pool. Our

results imply that both IPO auctions and equity crowdfunding/crowdinvesting should

include separate, non-price-setting tranches for the general public.

JEL Classification: C92, D02, D44, G02

I. Introduction

The method used for initial public offerings (IPOs) in the United States, known as book

building, has long been controversial because of the magnitude of average underpricing, the

variability of initial returns, and the conflicts of interest between issuers and underwriters.

Ritter (2011) discusses the CLAS controversies (Commissions for IPOs, Laddering, Analyst

conflicts of interest, and Spinning) and argues that agency problems between issuers and

underwriters are of first-order importance in explaining the magnitude of underpricing,

leading to questionable allocation practices that preclude participation by many investors.

1

We thank conference participants at the 2015 meeting of the Financial Management Association and seminar

participants at the University of Central Florida, Ivey Business School, Iowa State University, and the University of

Iowa for helpful comments.

1

In contrast, Sherman (2000) models the effect of the one price rule for U.S. IPOs when there is more than one

type of investor. If IPO shares must be underpriced for some investors (e.g., to attract attention or induce costly

evaluation), the one price rule means that other investors get excess returns unless the underwriter can somehow

recapture some of that excess. Thus, many of the allocation practices described by Ritter (2011)may occur without

affecting the underpricing level. See also Loughran and Ritter (2004) for discussion of the scandals and overall

trends in IPO underpricing, and Ritter and Welch (2002), Ljungqvist (2007), and Wilhelm (2005) for reviews of the

academic IPO literature.

The Journal of Financial Research Vol. XLII, No. 1 Pages 5–39 Spring 2019

DOI: 10.1111/jfir.12169

5

© 2019 The Southern Finance Association and the Southwestern Finance Association

Many have suggested auctions as an alternative, given the success of auctions in other

financial markets, such as government debt securities. After all, the surest way to prevent

agency conflicts is to take decisions out of the hands of the agent, which is what a standard

sealed bid auction does for IPO pricing—it uses preestablished rules to price the shares based

on bids, rather than giving discretion to the underwriter. IPO auctions offer transparent

allocation and pricing rules, and open access to retail investors.

Yet auction IPOs have not been popular in the United States. There have been a

total of 23 IPO auctions in the United States since the method was first introduced by the

investment bank WR Hambrecht in 1999. The auction method has been used for

prominent IPOs such as Google (2004) and Morningstar (2005); for tech/Internet IPOs

such as Andover.net (1999), Salon.com (1999), Bank of the Internet (2005), Netsuite

(2007), and Rackspace Hosting (2008); and for consumer-oriented, easily relatable

companies such as Ravenswood Winery (1999), Peet’s Coffee (2001), Overstock (2002),

and Truett-Hurst Winery (2013). Despite being used over time by a variety of companies,

the auction method has not taken off in the United States and has not been commonly

chosen by issuers. A public pool would involve setting aside a fixed number of shares for

retail investors, who could then participate in the IPO at the same price as investors in the

auction tranche. Our research indicates that a public pool would make the auction method

more attractive to both issuers and investors by increasing the expected proceeds for the

issuer and reducing risk for all involved.

The auction method has similarly struggled outside the United States. Sealed

bid auctions have been tried for IPOs in more than two dozen countries, but issuers

have consistently abandoned the auction method in favor of other alternatives.

2

The

experimentation with new IPO methods such as auctions began with the trend in the

1980s and 1990s toward privatization of formerly government-owned companies in

areas such as telecommunications. Auctions were often initially popular for IPOs,

but then issuers returned to their traditional method, the fixed-price public offer.

3

Most countries later adopted the book-building method that originated in the United

States.

4

Although it has been rare for IPO issuers to choose the auction method

voluntarily over the last decade, regulations in Vietnam, China, and Taiwan mandate

their use. The auction method has long been required for IPOs in Vietnam, a country that

is still transitioning away from a planned economy through the privatization of state-

owned enterprises. Sequential hybrid uniform price auctions have been required in China

2

Jagannathan, Jirnyi, and Sherman (2015) document the disappearance of auction IPOs and find that it was

not driven by regulatory changes that restricted their use (although in some cases, such as France and Japan, the

shift was made possible by regulations allowing greater use of other methods).

3

A public offer, also known as an open offer, a universal offer, or often simply as “the IPO method,”is one in

which the price is set in advance and the issuer/underwriter does not have discretion regarding individual

allocations. Typically, the process is open to all retail investors in the country on a more or less equal basis, and

balloting (a type of lottery) is used to allocate shares in case of oversubscription. A public offer is strictly an

allocation method, as opposed to a price-setting method.

4

Whereas Doidge, Karolyi, and Stulz (2013) show that U.S. IPOs have become less important around the

world over the past two decades, Jagannathan, Jirnyi, and Sherman (2015) show that the U.S. IPO method has

become dominant.

6 The Journal of Financial Research

since at least 2009 (see Cao et al. 2016). As of January 1, 2016, uniform price auctions are

required for large IPOs (more than NT$400 million in proceeds) in Taiwan.

Given the many well-established advantages of auctions and yet their consistent

unpopularity for IPOs, one explanation for this discrepancy is that the standard sealed bid

auction format is not as well suited to IPOs as to, say, Treasury securities. Price discovery

is less of an issue for Treasury securities, which are high-quality debt instruments that are

already trading in many forms (the when issued market, the off the run issue, etc.), with a

stable number of large dealers that bid for them regularly. IPO firms tend to be small,

young companies that have not been traded, and are likely to grow and change

dramatically and unpredictably in the next few years. They come from a variety of

industries and appeal to different sets of investors. Such different securities may require a

different auction format.

5

Thus, it is worth exploring ways to design auctions that are better suited to an

IPO environment. The first study to model this is by Biais and Faugeron-Crouzet (2002),

who explore the effects of the frequent IPO auction practice of setting the offer price

strictly below the market-clearing price (known as a “dirty”auction or as “leaving

something on the table”). We focus on another common feature of IPOs—the use of

hybrids, with separate retail tranches or public pools.

6

Our article offers experimental and

theoretical evidence that the inclusion of a public pool may improve the auction method

for IPOs by reducing risk for both bidders and the issuer.

We first examine the extent to which free riding, inaccurate pricing, and

variability in auction outcomes arise in a uniform price auction with endogenous costly

information acquisition. Next, we determine whether the introduction of a public pool

tranche can be used to reduce the problems that have been identified. We accomplish this

by conducting a laboratory experiment that incorporates key features of an IPO auction

environment, including endogenous entry and bidders of different bidding capacities. A

major advantage of experimental over “real-world”bidding data is that we know the

information set and budget constraints of each bidder and thus do not have to speculate

over, for example, whether a bidder is informed. The hypotheses we test are based on the

use of simulations to identify the symmetric Bayesian Nash equilibrium (SBNE) for both

pure and hybrid auctions.

7

5

The differences are even wider between Treasury securities and equity investments in early-stage private

companies. Thus, this research has implications for the equity crowdfunding made possible in the United States by

the Jumpstart Our Business Startups (JOBS) Act, including Regulation Aþofferings. These implications are

discussed briefly in the Conclusion.

6

These public pools originated from the traditional fixed-price public offer method, where investors could ask

for share allocations but were not involved in setting the offer price. Given the long tradition of allowing all retail

investors to order shares in each IPO, most countries that experimented with first the auction and then the book-

building method continued to allow retail investors to participate—in much the same way as before—through a

retail tranche or public pool. Thus, hybrid IPOs with a public pool are common internationally, although generally

the book-building rather than the auction method is used to set the offer price.

7

Throughout this article, we use the term “symmetric”to mean symmetric within each investor group, given

that our experiments each have four large and four small bidders, with different capacity constraints. Thus, we

consider a set of strategies to be symmetric even if large and small bidders follow different strategies, as long as

there are no strategy differences within each size group.

Hybrid IPO Auctions with a Public Pool 7