How independent is a 'qualified independent underwriter'?

Executive Summary

The main goal of this study is to provide statistical evidence about the effectiveness of a Qualified Independent Underwriter (QIU) to minimize conflicts of interest as defined by the Rule 2720 Distribution of Securities of Members and Affiliates--Conflicts of Interest. The important statistical evidence that we provide in this paper can be used by the Financial Industry Regulatory Authority (FINRA) to eliminate or modify Rule 2720 in order to avoid a biased IPO pricing function.

We measure the difference in underpricing between IPOs priced by QIUs, and equivalent IPOs with no conflicts of interest, using a matching group of IPOs carefully selected by IPO date, size, and industry sector.

We define the level of abnormal underpricing as the difference between the underpricing of each firm in the matching group and the underpricing of its corresponding firm in the sample group.

The cross-sectional variation of the level of abnormal underpricing is determined by an ordinary least square (OL5) regression model.

The results show that the level of underpricing is about 23.5 percent when an IPO is priced by a QIU, while this level increases up to 36.6 percent when the IPO is not priced by a QIU.

These results suggest that the conflicts of interest between issuer and underwriter are not mitigated by Rule 2720, and therefore this rule should be modified in order to provide more effective regulations to avoid the impact of conflicts of interest in the IPO pricing function.

Introduction

The purpose of Rule 2720 Distribution of Securities of Members and Affiliates--Conflicts of Interest, is to ensure that some aspects of the public offering of debt and equity securities, particularly the pricing function, are regulated when potential conflicts of interest exist. The rule requires the public disclosure of such potential conflicts of interest, as well as the appointment of a Qualified Independent Underwriter who will be in charge of the pricing function. The participation of a QIU in an IPO is intended to protect issuers and investors by mitigating inherent conflicts of interest. The best interest of an issuer is to maximize the proceeds of the IPO, while investors' best interest is to maximize the return on their investment through highly underpriced IPOs.

Collusion in the underwriting business involves investment banks cooperating or working together, instead of competing with one another. By collaborating with each other, underwriters can alter the price of an IPO and take advantage of issuers and/ or investors. Barondes reports that "recent events are replete with stories of fraudulent or opportunistic behavior in the IPO-pricing process ..." (Barondes 2005a, p. 2). Conclusive evidence about collusive behavior is extremely hard to provide since participants are not imprudent enough to put arrangements to collude on paper. Collusion is more likely to happen when there are a few participants in the market that offer a product or service to a large number of consumers. Fu and Li (2007) report that the IPO market is highly concentrated, since the top four underwriters comprise more than 50 percent of the IPO business. They conclude that the high market concentration in the underwriting business is both a signal and a contribution to the collusion. Cetorelli et al. (2007) also find that since 1990 the IPO and seasoned equity markets have experienced an increase in the level of market concentration.

The first work to provide evidence about possible underwriter collusion is Chen and Ritter (2000), who study the gross spreads of IPOs issued between 1985 and 1998. They find that more than 90 percent of IPOs issued between 1995 and 1998 in the $20 to $80 million range paid exactly seven percent of the proceeds to their underwriters. They explain these results as possible implicit collusion among investment banks. These findings led to the antitrust division of the U.S. Department of Justice to conduct an investigation about possible collusion among investment banks. However, other academic works provide evidence that the Chen and Ritter's collusion theory is incorrect (see Torstila, 2003, and Hansen, 2001).

Barondes (2005b), like this article, also provides significant evidence against the effectiveness of Rule 2720 to mitigate conflicts of interest. However, Barondes (2005b) clarifies that his article does not attempt to determine whether IPO pricing involving QIU is correct compared to those IPOs available otherwise in the market. This is precisely the original contribution of this article compared to that of Barondes (2005b). By using a matching group of IPOs carefully selected by IPO date, size, and industry sector, this article tries to measure the difference in underpricing between IPOs priced by QIUs, and equivalent IPOs with no conflicts of interest. Barondes (2005b) states that " ... there is a potential concern that may suggest it is desirable to confirm the results by an alternative technique. The problem is that we would like to be able to examine pairs of IPOs that are identical but for the presence of a conflict of interest. We cannot do that, of course..." (2005b, p. 893.)

We agree with Barondes in the sense that identical IPOs do not exist; however, we think that very similar IPOs in terms of the three variables mentioned above can be obtained and used to confirm his results from a different methodological perspective. Barondes (2005b) highlights that the IPO underpricing was extreme during late 1990s, which is the period of time considered in his study. He admits that his results may not apply to other time periods, where they may have been identified with different types of IPO markets. Barondes' sample consists of 1188 IPOs from 1997 through 2000, where just 133 IPOs were priced by QIUs. Our sample consists of 190 IPOs priced by QIUs (plus a matching group of 190 IPOs with no conflicts of interest) from 1996 through 2003. Therefore, our study provides evidence of the effectiveness of Rule 2720 in both cold and hot IPO markets.

The rest of this paper is organized as follows:

* section 1 summarizes the relevant literature review on this topic;

* section 2 explains a brief overview about Rule 2720 and describes some testable hypotheses;

* section 3 provides our methodology and proposed model;

* section 4 describes the sample with some descriptive statistics;

* section 5 explains the empirical results; and

* section 6 summarizes the major findings.

1. Literature Review

The study of the effectiveness of the Securities and Exchange Commission's Rule 2720 Distribution of Securities of Members and Affiliates--Conflicts of Interest has only been considered in three academic works. Barondes (2005b) finds different levels of underpricing depending on the type of conflict of interest that a QIU is intended to mitigate. When 10 percent or more of the issuer's preferred stocks or subordinated debt is owned by any investment bank participating in the IPO, or when more than 10 percent of the proceeds are intended to be paid to one of the IPO's underwriters, there is a negative relationship between the underpricing and the presence of a QIU of -27 and -15 percent respectively. When 10 percent or more of the issuer's common stocks or subordinated debt is owned by any investment bank participating in the IPO, there is a positive relationship between the underpricing and the presence of a QIU of 15 percent. He concludes that the regulatory environment designed to make the IPO pricing function invariant to the presence of conflicts of interest as defined by Rule 2720, is a clear failure.

Li and Masulis (2004) study the IPO pricing function when investment banks are both underwriters and prior venture investors in IPO issuers. They propose the issuer alignment hypothesis where underwriters with prior equity investment in issuers have an incentive to set high offer prices because otherwise their equity investment would be diluted. They find that prior underwriter equity ownership significantly decreases IPO underpricing. The reduction in underpricing is stronger in the case of lead underwriter ownership. They also examine the effects of NASD Rule 2720 on IPO underpricing. They suggest that there might be conflicts of interest between IPO investors and underwriters acting as QIUs since these investment banks want to be included in future syndicates with the same lead underwriter. They find that "... When a QIU indicator is added as a new control variable in the statistical model... we find that it is statistically insignificant. This evidence suggests that the QIU mechanism is ineffective at protecting IPO investors..." (p. 19). This inconsistency in Li and Masulis's (2004) results is also observed by Barondes, who asserts that " ... (Li and Masulis) conclude, on the basis of the insignificance of the variable, that 'this evidence suggests that the QIU mechanism is ineffective in protecting IPO investors'. ... They do not explain the rationale for that conclusion, and it does not appear to follow." (2005b, p. 880)

Barondes (2005a) also analyzes investment bank conflicts of interest in IPO pricing, through the study of pre-lPO price adjustments of IPOs priced by QIU. He finds relatively lower pre-lPO price adjustment when more than 10 percent of the proceeds are intended to be paid to participating investment banks. He asserts that the primary contribution of his article is to rebut Li and Masulis's conclusions.

One of the first articles to address the conflict of interest between IPO issuers and underwriters is Baron and Holmstrom (1980). They propose a model where the uninformed issuer delegates the IPO pricing decision to the investment bank. The investment bank or underwriter can make a better IPO pricing decision based on its superior information. This delegation results in a price that would not be the same if the issuer had the same information as the investment bank. Muscarella and Vetsuypens (1989) tested Baron's model. They hypothesized that if this model is correct then they should find no underpricing in their sample of self-underwritten IPOs. However, their empirical results provide evidence that self-underwritten IPOs show a level of underpricing of about the same as that of IPOs underwritten by investment banks.

Print Get the Mag Weekly Updates