The effect of the trading system on the underpricing of initial public offerings.

Two major trading systems in common stocks in the United States are (i) the exchange auction markets of the New York Stock Exchange (NYSE) and the American Stock Exchange (AMEX), and (ii) the negotiated dealer markets of the National Association of Securities Dealers' Automated Quotations (NASDAQ), National Market System (NMS) and non-NMS. These trading systems are further distinguished by their initial and continued standards for stock trading (i.e., listing).(1) The objective of this study is twofold. First, we examine whether the underpricing previously documented for initial public offerings (IPOs) on NASDAQ also applies to IPOs on the AMEX and NYSE. Second, we examine the effects of these trading systems on the pricing of IPOs.

A firm that is going public is surrounded by substantial informational asymmetry and the extant theoretical and empirical literature suggests this leads to the underpricing of IPOs (Baron |5~, Beatty and Ritter |6~, Grinblatt and Hwang |12~, Rock |25~, Smith |26~, Tinic |27~, and Welch |31~). We advance a trading system certification proposition which posits that the initial and continued listing standards imposed by a trading system provide investors with reliable information about the quality of new issues, reduce uncertainty about their prospects, and thereby lower the expected underpricing of IPOs.

Using a sample of IPOs which went public on the various systems over the period 1983-1987, we show that IPOs on all four trading systems display significant underpricing, on average. We also provide support for the certification proposition. After controlling for other factors previously shown to be associated with the underpricing of IPOs, namely, offer size, ownership retention, age of the firm, standard deviation of returns in the aftermarket, prestige of the underwriter, and auditor reputation, we show that NYSE, AMEX and NASDAQ/NMS IPOs are associated with significantly lower underpricing than NASDAQ/non-NMS IPOs. Further analysis indicates that no significant differences in underpricing exist between the NYSE, AMEX and NASDAQ/NMS trading systems, after controlling for other ex ante uncertainty proxies.

The rest of the paper is organized as follows. In Section I, we examine the role of initial and continued listing standards in certifying the quality of new issues and develop our certification proposition. We present the empirical results in Section II, while Section III contains several robustness checks on our main results. Section IV sets forth a summary of our findings and conclusions.

I. Trading Systems and the Underpricing of IPOs

In this section, we develop a certification proposition regarding the impact of trading systems on the underpricing of IPOs. We begin by summarizing the listing standards of the four market systems and then discuss how these standards mitigate uncertainty about the quality of new issues through certification.

A. Listing Standards

An important distinguishing characteristic of a trading system is its listing standards. There are two types of listing standards: quantitative and qualitative. Also, there are initial as well as continued listing standards. Exhibit 1 summarizes the minimum quantitative initial and continued listing standards for the NYSE, AMEX, NASDAQ/NMS, and NASDAQ/non-NMS markets. Clearly, there is wide variation in the requirements related to income, public ownership, firm size, and stockholder equity across the trading systems. Notably, size, distribution of ownership and earnings standards are substantially higher for the NYSE, somewhat comparable for the AMEX and NASDAQ/NMS, and least stringent for the NASDAQ/non-NMS.

In addition, the listing agreement includes numerous qualitative standards designed to protect shareholders from agency problems ensuing from the separation of ownership and management of a firm. These standards are meant to (i) ensure timely disclosure of information that may affect security value or influence investment decisions, (ii) encourage and enforce certain business practices aimed at maintaining sound standards of corporate responsibility, integrity and accountability to shareholders, and (iii) provide the trading system with timely information to facilitate the maintenance of a fair and orderly market in the firm's securities.

Major issues covered by the qualitative standards of the NYSE are: the number of outside directors, the representation of independent directors on the audit committee, the review of related party transactions, the quorum required for shareholder meetings to ensure a representative vote, and the shareholder approval policy for securities issued to directors, officers, or key employees (see New York Stock Exchange |21~). These qualitative standards are also representative of the corporate governance rules adopted by the AMEX.(2) In early 1985, the NASD board of governors developed similar rules for the NMS securities. These rules were approved by the SEC in June 1987 and implemented in February 1989. The NASD corporate advisory board observed that most NMS companies already complied with the new corporate governance standards.(3)

Clearly there is some duplication between these qualitative standards and the Securities Act of 1933 and Securities Exchange Act of 1934, which apply to all publicly traded companies. Several provisions of the corporate governance standards, however, go beyond the SEC rules. Partly as a result of these extended provisions, securities listed on the NYSE and AMEX are automatically exempt from blue-sky regulations in most states. Obtaining a similar exemption was a major objective of the new NASDAQ/NMS rules. Further, to avoid these high standards, many firms that are quantitatively eligible for listing on the NYSE and AMEX may elect to continue trading on NASDAQ. A 1985 study by Wall |29~ relative to this possibility indicated that more than 600 NASDAQ companies are eligible for listing on the NYSE and 1,600 are eligible for listing on the AMEX, although of course, there are other reasons, such as direct costs or the existence of an active market in their stock on NASDAQ, which may impact the trading system decision.

TABULAR DATA OMITTED

B. Certification of Firm Quality

A trading system(4) seeks to maintain high listing standards not only to maintain an orderly market for securities, but also to develop a reputation as a leading marketplace for securities trading. In fact, the NYSE states,

A listing on the NYSE is internationally recognized as signifying that a publicly owned corporation has achieved maturity and front-rank status in its industry -- in terms of assets, earnings, and shareholder interest and acceptance. Indeed, the Exchange's listing standards are designed to assure that every domestic or non-U.S. company whose shares are admitted to trading in the exchange market merits that recognition. |21, p. 11~

Commenting on its listing requirements, the NYSE observes,

In determining eligibility for listing, particular attention is given to such qualifications as: (1) the degree of national interest in the company; (2) its relative position and stability in the industry; and (3) whether it is engaged in an expanding industry, with prospects of at least maintaining its relative position. |22, p. 22~.

A privately held firm intending to go public is characterized by a great deal of informational asymmetry about its prospects between insiders (the firm) and outside investors. By seeking to list a new public offering, the management of the firm asserts that it meets the initial listing standards, that it expects to satisfy the continued quantitative listing standards, and that it is committed to honoring the trading system's corporate governance standards.(5) This managerial signal derives credibility from several factors. First, the firm expends resources on preparing the initial listing forms and payment of the initial listing fees and subjects itself to recurring annual listing fees. Second, by agreeing to abide by the system's corporate governance standards, the firm incurs higher shareholder servicing costs and potential managerial inconveniences and inflexibilities. Third, the firm exposes itself to the risk of being delisted if it fails to maintain the continued listing standards.

More importantly, listing carries with it the approval and reputation of the trading system because the trading system effectively certifies that the new issue meets its initial quantitative and qualitative admission standards.(6) The latter mitigates the agency problems confronting investors. The importance trading systems attach to this certification is illustrated by the AMEX which states,

Since new issues may pose investment and trading risks not normally encountered in securities of a company which is already publicly traded, the Exchange has followed the practice of accepting an IPO for trading only when it is demonstrated that the issue will exceed all the numerical guidelines for original listing. (emphasis added) |3, p. 11~

A trading system has incentive to carefully monitor these initial and continued listing standards because (i) the listing standards are public information, (ii) there is keen competition for listing securities among the NYSE, AMEX, and NASDAQ/NMS as evidenced by the promotional literature distributed by the trading systems, and (iii) the trading system has nonsalvageable reputation capital at stake.

C. Informational Asymmetry and Underpricing of IPOs

An important consequence of informational asymmetry is the equilibrium underpricing of IPOs. Rock |25~ views this situation as a winner's curse problem. In his model, there exist two groups of investors: the informed investors who possess superior information, and uninformed investors (including the firm as well as other investors). To ensure continued liquidity of the IPO market, new issues must be underpriced to compensate uninformed investors for this adverse selection risk.

We use Beatty and Ritter's |6~ framework to test the impact of listing on IPO underpricing in the Rock framework.(7) They show that as ex ante uncertainty about the value of a new issue increases so does the expected underpricing. We have argued that, through their listing standards, trading systems certify the quality of a new issue which partially resolves the informational asymmetry between insiders and outside investors. Therefore, given a choice between two new issues scheduled to be listed on two alternative trading systems with differential listing standards, we argue that investors will demand less underpricing from the issue associated with the trading system with higher listing standards.(8) This leads to our certification proposition:

Proposition: Higher listing standards reduce the expected underpricing of an IPO by reducing ex ante uncertainty about its value.

It is important to note that certification could exist at several levels. For example, there could be a strict rank ordering from NYSE to AMEX to NASDAQ/NMS to NASDAQ/non-NMS. Alternatively, there could be relatively little difference between the NYSE and AMEX, but considerable difference between the exchanges and NASDAQ. Finally, the major certification difference could be between the exchanges and NASDAQ/NMS on the one hand and the NASDAQ/non-NMS IPOs on the other.

These three issues will be examined empirically in the following section. Our prior analysis suggests that many of the qualitative standards are comparable across the NYSE, AMEX, and NASDAQ/NMS,(9) because of NASDAQ's attempts to raise standards in order to gain blue-sky exemption for NMS-listed securities. If this is true, the largest difference may lie between the group comprising the exchanges plus the NASDAQ/NMS IPOs and the NASDAQ/non-NMS IPOs.

II. Empirical Analysis

A. Data

The NYSE developed special procedures in 1983 to enable firms to list concurrently with their initial public offering of common stock. Accordingly, our study begins with January 1983 and covers IPOs on the NYSE, AMEX and NASDAQ markets through December 1987. The sample includes all firm commitment IPOs on the NYSE and AMEX for which a closing price on the first day of trading was available. This resulted in a sample of 95 IPOs that were listed on the NYSE and 50 IPOs on the AMEX. Over the same five-year period, we found 158 NMS and 824 non-NMS firm commitment IPOs that were listed on the NASDAQ system,(10) with a first quoted bid of at least a dollar.(11) All price and return data were obtained from the CRSP NYSE/AMEX and NASDAQ tapes. Offer prices on the IPOs were obtained from Going Public: The IPO Reporter and Investment Dealers Digest.

Several of the IPOs in the exchange samples were closed-end funds, real estate investment trusts, master limited partnerships, or reorganizations (i.e., spinoffs or restructurings). Because IPOs such as these may have particular characteristics different from IPOs in general (e.g., Weiss |30~ has shown little underpricing in closed-end fund IPOs), the presence of these in the exchange sample may bias our results. Accordingly, we remove all IPOs in the above four classes from our samples of NYSE and AMEX IPOs.(12)

B. Underpricing

Exhibit 2 presents summary statistics on the underpricing of our sample of IPOs. We use initial return, IR, as a measure of underpricing, where IR is defined as the excess (in percent) of the first day's closing price over the offer price (not adjusted for market movements). The average first day returns (IR) for the NYSE, AMEX, NASDAQ/NMS, and NASDAQ/non-NMS IPOs are 4.82%, 2.16%, 5.56%, and 10.41%, respectively. Each of these average first day returns is significantly different from zero (t-values of 3.91, 2.24, 7.04, and 17.39, respectively) indicating that significant underpricing, on average, exists in each of the trading systems examined. Our results for the NYSE and AMEX IPOs are consistent with those of Ibbotson |15~ and confirm that IPOs on the exchanges TABULAR DATA OMITTED display an underpricing phenomenon similar to that documented, for example, by Ritter |23~ for NASDAQ stocks. In addition, the mean initial returns from NYSE and AMEX IPOs are smaller than those from the NASDAQ/NMS and NASDAQ/non-NMS IPOs, which is consistent with our expectation, although it does not validate our first proposition, since factors such as risk and size are not controlled for in these initial tests.

Our proposition makes a stronger claim because it asserts that the lower underpricing of the NYSE, AMEX and NASDAQ/NMS IPOs is attributable to their listing standards which reduce the ex ante uncertainty about the value of a new issue. To test this proposition, we control for six accepted proxies for ex ante uncertainty: (i) OR, ownership retention (Grinblatt and Hwang |12~ and Leland and Pyle |16~); (ii) AGE, the age of the firm on the date of the initial public offering (Carter and Manaster |8~); (iii) RGP, the reciprocal of gross proceeds (Beatty and Ritter |6~); (iv) IB, the prestige of the investment banker (Carter and Manaster |8~); (v) AUD, the reputation of the auditor (Balvers, McDonald and Miller |4~); and (vi) |Sigma~, the standard deviation of daily aftermarket returns estimated over the first 20 days in the aftermarket (Ritter |23~).(13)

Our test procedure is to fit the following cross-sectional regression model:

|IR.sub.i~ = ||Alpha~.sub.0~ + ||Alpha~.sub.1~|OR.sub.i~ + ||Alpha~.sub.2~|AGE.sub.i~ + ||Alpha~.sub.3~|RGP.sub.i~ + ||Alpha~.sub.4~|IB.sub.i~ + ||Alpha~.sub.5~|AUD.sub.i~ + ||Alpha~.sub.6~||Sigma~.sub.i~ + ||Alpha~.sub.7~(|IB.sub.i~ |center dot~ |AUD.sub.i~) + ||Alpha~.sub.8~|E.sub.1i~ + ||Alpha~.sub.9~|E.sub.2i~ + ||Alpha~.sub.10~|E.sub.3i~ + |e.sub.i~. (1)

where |OR.sub.i~, ownership retention, is the percentage of shares retained by the original owner (i.e., not offered in the initial public offering); |AGE.sub.i~ is the number of years the firm was incorporated prior to the initial public offering (obtained from Moody's Manuals and Ward Business Directory); |RGP.sub.i~ is the reciprocal of gross proceeds from the offering, defined as the number of shares offered times the offering price; |IB.sub.i~ equals one if the investment banker was in the bulge bracket as defined by Hayes |13~, i.e., the top five investment banking firms in the United States, and zero otherwise; |AUD.sub.i~ equals one if the auditor was a big-8 (now big-6) auditor, and zero otherwise; ||Sigma~.sub.i~ is the standard deviation of the first 20 daily returns (as calculated from the average of the closing bid and ask quotes each day)(14) in the aftermarket; and |E.sub.1~, |E.sub.2~ and |E.sub.3~ are binary variables which assume a value of one for the NYSE, AMEX and NASDAQ/NMS IPOs, respectively, and are zero otherwise. The interaction between investment banker and auditor is included following Balvers, McDonald, and Miller |4~. Previous research suggests ||Alpha~.sub.1~ |is less than~ 0, ||Alpha~.sub.2~ |is less than~ 0, ||Alpha~.sub.3~ |is greater than~ 0, ||Alpha~.sub.4~ |is less than~ 0, ||Alpha~.sub.5~ |is less than~ 0, ||Alpha~.sub.6~ |is greater than~ 0 and ||Alpha~.sub.7~ |is greater than~ 0. Our proposition implies negative coefficients associated with |E.sub.1~, |E.sub.2~ and |E.sub.3~, indicating that underpricing is lower on these market centers after controlling for other accepted measures of ex ante uncertainty.

Before reporting our regression results, we provide summary statistics for each of the six proxies in Exhibit 3. The summary statistics for the proxies show some differences across the trading systems. For example, gross proceeds is largest by far for the NYSE IPOs, followed by AMEX, NASDAQ/NMS and NASDAQ/non-NMS in decreasing order of magnitude. Similar comments apply to the firm value variable. About 47% of IPOs on the NYSE are taken public by an investment banker in the "bulge bracket" -- i.e., one of the five largest investment banks. TABULAR DATA OMITTED The proportion declines to approximately 12% and 24% for AMEX and NASDAQ/NMS IPOs, respectively, while less than 10% of IPOs on NASDAQ/non-NMS are associated with a bulge bracket investment banker. Finally, over 80% of IPOs on all four trading systems are associated with a big-8 auditor.

We report the ordinary least squares regression estimates of Equation (1) in the first column of Exhibit 4. The overall regression is significant at the one percent level. In addition, the signs for all seven variables included to capture ex ante uncertainty have the predicted signs, and all, except for ownership retention, are statistically significant at the five percent level.

Coefficients ||Alpha~.sub.8~ through ||Alpha~.sub.10~ show the extent to which the initial NYSE, AMEX and NASDAQ/NMS IPO returns differ from their NASDAQ/non-NMS counterparts after controlling for the six proxies for ex ante uncertainty. If the proposition that listing standards serve as a signal of investment and managerial quality of IPOs has any merit, we should expect that, at the minimum, the NYSE, AMEX and NASDAQ/NMS IPOs have initial returns that are lower than those of the NASDAQ/non-NMS IPOs. This is reflected in the following joint null (|H.sub.0~) and alternative (|H.sub.A~) hypotheses:

|H.sub.01~: ||Alpha~.sub.8~ = ||Alpha~.sub.9~ = ||Alpha~.sub.10~ = 0,

|H.sub.A~: not all ||Alpha~.sub.k~ equal 0, k = 8, 9 and 10.

The F-statistic for the joint hypothesis is 3.98 (with 3 and 1,067 degrees of freedom) and is statistically significant at the one percent level. Further, the t-statistics for the individual coefficients indicate that the estimates |Mathematical Expression Omitted~ and |Mathematical Expression Omitted~ are significantly negative, while |Mathematical Expression Omitted~ is negative but not statistically significant.(15)

Having identified that initial IPO returns are significantly lower on the NYSE, AMEX, and NASDAQ/NMS relative to NASDAQ/non-NMS stocks, we next examine whether the degree of underpricing varies across the NYSE, AMEX and NASDAQ/NMS IPOs. This is reflected in the following joint null hypothesis:

|H.sub.02~: ||Alpha~.sub.8~ = ||Alpha~.sub.9~ = ||Alpha~.sub.10~.

The F-statistic for this joint hypothesis is 1.06, which is not significant at conventional levels. Consequently, we conclude that there is no significant difference in the underpricing of IPOs listed on the NYSE, AMEX, and NASDAQ/NMS after controlling for ex ante uncertainty and hence that the major certification role is played by the exchanges or NMS for NASDAQ stocks versus non-NMS TABULAR DATA OMITTED stocks. Although our data do not allow us to determine why certification only differs across the exchanges and NMS versus non-NMS dimension and not between the exchanges and NASDAQ/NMS IPOs, we suggest it may be because the NASDAQ/NMS is the prestige segment of the NASDAQ market (which itself is the upper echelon of the total OTC market). Put another way, there are three segments of the OTC market: (i) NASDAQ/NMS, (ii) NASDAQ/non-NMS, and (iii) non-NASDAQ. As a result, listing on the NMS segment of NASDAQ is certification for OTC stocks and makes these stocks similar to NYSE stocks in the eyes of IPO investors. There is support for this higher status based on the differential listing requirements for the two NASDAQ groups as shown in Exhibit 1 (especially net income, shares publicly held, and market value of shares). There is also empirical evidence of the prestige of NASDAQ/NMS from the sample in Exhibit 3. Specifically, only about 16% of the IPOs that were traded on the total NASDAQ system were admitted to the NMS segment of NASDAQ (158 of 982). Also, 24% of the NMS stocks had prestigious investment bankers which was a larger percentage than for the AMEX sample (12%), and a much higher percentage than for the non-NMS sample (10%).

TABULAR DATA OMITTED

III. Robustness Checks

In this section, we present several checks to demonstrate that our results are not sensitive to model specification. These checks include a weighted least squares to control for possible heteroskedasticity, inclusion of price in the regression to control for transaction costs (Chalk and Peavy |9~), and alternative proxies to the reciprocal of gross proceeds (namely log (gross proceeds) and log (market value)). The results of all robustness checks are summarized in columns 2 to 6 of Exhibit 4. An ordinary least squares regression of initial returns on the ex ante proxy explanatory variables may be characterized by heteroskedasticity. Since firm value and ex ante uncertainty tend to vary inversely, we multiplied all variables by the log of market value (market capitalization at the close of the first day of trading). This weighting procedure is expected to produce homoskedastic disturbances and is in the spirit of Beatty and Ritter |6~. The results are presented in the second column of Exhibit 4 and are qualitatively identical to those in the first column in terms of the signs and significance of the coefficients.

Chalk and Peavy |9~ suggest inclusion of price in the regression model to control for the effects of transaction costs. In particular, Chalk and Peavy find a negative relationship between price and the level of underpricing. Inclusion of price in Equation (1) results in the OLS estimates reported in column 3 of Exhibit 4. Once again, the coefficients are very similar in terms of both signs and significance to those in the first column. The coefficient of price is significant but positive, possibly as a result of the partial adjustment phenomenon (Hanley |14~). Nevertheless, our major result, namely the significant difference between the NASDAQ/non-NMS IPOs and the other IPOs, is not altered.

In Equation (1), we used the reciprocal of gross proceeds as a proxy for the size of the offering. To test the robustness of the results to this specification we repeat the analysis using two alternative size proxies, the log of gross proceeds and the log of market value (FS, defined as number of shares outstanding times price at the close of the first day of trading). Once again, the results (columns 4 and 5 of Exhibit 4), which indicate a significant difference between NASDAQ/non-NMS IPOs and the other IPOs, are essentially unaltered.

Finally, we examine the distribution of offerings by trading systems over the calendar years 1983 to 1987. The results are presented in Exhibit 5, which also gives the mean underpricing in each year for each system. Clearly, the majority of the offerings on the exchanges occurred in 1986 and 1987. Similarly, all but one of the NASDAQ/NMS offerings occurred in these two years. We, therefore, reestimated the regression model using only data for 1986 and 1987. The results are presented in the final column of Exhibit 4 and are qualitatively unchanged from those in column 1.

The use of nominal gross proceeds in the regressions may introduce measurement error as the level of the market was much higher in August 1987 than in January 1983. The results in the final column of Exhibit 4, which are based on only 1986 and 1987 data, provide some evidence that our results are not being driven by the measurement error problem. The above results suggest that the evidence in favor of our certification hypotheses is not sensitive to model specification. Indeed, the results show considerable consistency, with the coefficients of most variables being similar across alternative specifications. While some differences emerge, these are probably due to multicollinearity in the data and do not affect our major conclusion, namely that initial returns differ significantly between NASDAQ/non-NMS IPOs and those on the other trading systems (i.e., the exchanges and NASDAQ/NMS).

IV. Summary and Conclusion

We present results showing that the previously documented average underpricing of IPOs on the NASDAQ system extends to both NYSE and AMEX IPOs. More specifically, we find that our sample of NYSE IPOs are, on average, underpriced by 4.82%, while our sample of AMEX IPOs have an average underpricing of 2.16%. This contrasts with the 5.56% and 10.41%, respectively, for NASDAQ/NMS and NASDAQ/non-NMS IPOs, on average, over the same time period.

We also advance the proposition that, through their quantitative and qualitative initial and continued listing standards, trading systems certify the quality of IPOs, including managerial quality narrowly defined along its agency dimension. The higher listing standards of the NYSE, AMEX and NASDAQ/NMS relative to the NASDAQ/non-NMS reduce the ex ante uncertainty about the value of an IPO and thus reduce its expected underpricing.

Our empirical results support this proposition with the NYSE, AMEX, and NASDAQ/NMS IPOs having significantly lower underpricing than non-NMS IPOs after controlling for several other proxies for ex ante uncertainty. However, while the average underpricing for NYSE and AMEX IPOs was less than that of NASDAQ/NMS IPOs, these differences in underpricing were not statistically significant once other sources of ex ante uncertainty were controlled.

1 We use the phrase "listing standards" in a broad sense to refer to admission standards for trading not only on the exchanges but also on the NASDAQ. For a discussion of listing standards, see American Stock Exchange |1~, National Association of Securities Dealers |19~ and |20~, and New York Stock Exchange |22~.

2 For a discussion on the difference in corporate governance standards between the NYSE and AMEX, see American Stock Exchange |2~ and |3~. Also note that the listing standards are only guidelines, as evidenced by some listed firms having negative book values of shareholder equity.

3 See National Association of Securities Dealers |19, p. 35~.

4 We use the term "trading system" throughout the paper because we wish to compare both the specialist-based exchanges and the competitive dealer NASDAQ system.

5 Arguments along these lines are advanced in the context of listing of seasoned offerings by Ying, Lewellen, Schlarbaum, and Lease |32~.

6 On the application of reputational signaling to financial markets, see Beatty and Ritter |6~, Booth and Smith |7~, DeAngelo |10~, Easterbrook |11~, Mayers and Smith |17~, and Wakeman |28~.

7 In Rock's model, the firm is uninformed. In our setting, we assume the firm fully conveys its private information through information releases (for example, the prospectus) and its choice of the trading system. Thereafter, the firm is assumed to be in the same position as other uninformed investors.

8 The predicted effects of trading systems are also consistent with other theories of underpricing of IPOs. In the models of Grinblatt and Hwang |12~ and Welch |31~, underpricing serves as a signal of inside information about firm quality. Since certification through listing standards also signals firm quality both in its economic and agency dimensions, it serves as an alternative signal to underpricing. This implies that firms listing on trading systems with higher standards need to underprice less. In Tinic's |27~ interpretation, our argument suggests that higher listing standards reduce underpricing by mitigating the expected adverse legal consequences and damages to the reputation of investment bankers. Finally, Merton |18~ predicts that a larger investor base leads to wider recognition of the firm. Assuming higher listing standards make a security more attractive to both individual and institutional investors, these standards promote broader ownership of the firm, which reduces the severity of informational asymmetry and thus reduces the expected level of underpricing of IPOs.

9 Recall that although the NASDAQ/NMS implemented the qualitative standards in early 1989, many NMS firms already had business practices in place that were consistent with the new standards.

10 Best efforts IPOs were excluded because none of our sample of NYSE or AMEX IPOs were best efforts offerings. Ritter |24~ has documented that best efforts IPOs on NASDAQ have different initial return behavior than firm commitment offerings. To make our samples of NASDAQ IPOs as similar as possible to our NYSE and AMEX samples and to avoid a bias which may be introduced by including best efforts offerings in some groups and not others, it was decided to limit the study to firm commitment offerings. Similarly, all unit offerings (i.e., offerings with warrants attached) were excluded.

11 We discarded IPOs with a first quoted bid of less than a dollar in order to mitigate the bias in returns induced by discrete price quotes.

12 Although the impact on the size of the NYSE sample was substantial (reduction from 95 to 55 IPOs), the results are essentially unaltered when all IPOs are included.

13 Within the literature, numerous other proxies have been used. However, there is no consensus as to which is most appropriate and hence we employ six widely used and readily available proxies. We do not claim that these are either the best or the only possible proxies but believe that they adequately capture ex ante uncertainty.

14 Closing bid and ask quotes for AMEX and NYSE stocks were provided by the American Stock Exchange.

15 The lack of significance on ||Alpha~.sub.8~ may be caused by lack of statistical power due to the small sample size.

References

1. American Stock Exchange, Inc., Requirements for Original Listing, New York, 1987.

2. American Stock Exchange, Inc., Disclosure Policies, New York, 1987.

3. American Stock Exchange, Inc., Resource Prospectus, New York, 1988.

4. R. Balvers, B. McDonald, and R.E. Miller, "Underpricing of New Issues and the Choice of Auditor as a Signal of Investment Banker Reputation," Accounting Review (October 1988), pp. 605-622.

5. D.P. Baron, "A Model of the Demand for Investment Banking, Advising and Distribution Services for New Issues," Journal of Finance (September 1982), pp. 955-976.

6. R.P. Beatty and J.R. Ritter, "Investment Banking, Reputation and the Underpricing of Initial Public Offerings," Journal of Financial Economics (March 1986), pp. 213-232.

7. J.R. Booth and R.L. Smith, III, "Capital Raising, Underwriting and the Certification Hypothesis," Journal of Financial Economics (March 1986), pp. 261-281.

8. R. Carter and S. Manaster, "Initial Public Offerings and Underwriter Reputation," Journal of Finance (September 1990), pp. 1045-1067.

9. A.J. Chalk and J.W. Peavy, "Initial Public Offerings: Daily Returns, Offering Types and the Price Effect," Financial Analysts Journal (September-October 1987), pp. 65-69.

10. L. DeAngelo, "Auditor Independence, 'Low Balling,' and Disclosure Regulation," Journal of Accounting and Economics (August 1981), pp. 113-127.

11. F. Easterbrook, "Two Agency-Cost Explanations of Dividends," American Economic Review (September 1986), pp. 650-660.

12. M. Grinblatt and C.Y. Hwang, "Signalling and the Pricing of New Issues," Journal of Finance (June 1989) pp. 393-420.

13. S. Hayes, "Investment Banking: Power Structure In Flux," Harvard Business Review (March/April 1971), pp. 136-152.

14. K. Hanley, "Underpricing of Initial Public Offerings and the Partial Adjustment Phenomenon," Working Paper, University of Michigan, 1992.

15. R.G. Ibbotson, "Price Performance of Common Stock New Issues," Journal of Financial Economics (September 1975), pp. 235-272.

16. H. Leland and D. Pyle, "Information Asymmetries, Financial Structure, and Financial Intermediation," Journal of Finance (June 1977) pp. 371-387.

17. D. Mayers and C. Smith, "On the Corporate Demand for Insurance," Journal of Business (April 1982), pp. 281-296.

18. R.C. Merton, "A Simple Model of Capital Market Equilibrium with Incomplete Information," Journal of Finance (July 1987), pp. 483-510.

19. National Association of Securities Dealers, Inc., Reference Guide for NASDAQ Companies, New York, 1988.

20. National Association of Securities Dealers, Inc., Fact Book, New York, 1988.

21. New York Stock Exchange, Inc., Listed Company Manual, New York, 1983.

22. New York Stock Exchange, Inc., Fact Book, New York, 1988.

23. J. Ritter, "The 'Hot Issue' Market of 1980," Journal of Business (April 1986), pp. 215-240.

24. J. Ritter, "The Costs of Going Public," Journal of Financial Economics (December 1987), pp. 269-289.

25. K. Rock, "Why New Issues are Underpriced," Journal of Financial Economics (March 1986), pp. 187-212.

26. C.W. Smith, "Investment Banking and the Capital Acquisition Process," Journal of Financial Economics (March 1986), pp. 3-29.

27. S.M. Tinic, "Anatomy of Initial Public Offerings of Common Stock," Journal of Finance (September 1988), pp. 789-822.

28. L. Wakeman, "The Real Function of Bond Rating Agencies," Chase Financial Quarterly (Fall 1981), pp. 18-26.

29. J.T. Wall, "The Competitive Environment of the Securities Market," in Market-Making and the Changing Structure of the Securities Industries, Y. Amihud, T.S.Y. Ho, and R.A. Schwartz (eds.), Lexington Books, 1985, pp. 131-144.

30. K. Weiss, "The Post-Offering Price Performance of Closed-End Funds," Financial Management (Autumn 1989), pp. 57-67.

31. I. Welch, "Seasonal Offerings, Imitation Costs, and the Underpricing of Initial Public Offerings," Journal of Finance (June 1992), pp. 695-732.

32. L.K.W. Ying, W.G. Lewellen, G.G. Schlarbaum, and R.C. Lease, "Stock Exchange Listings and Securities Returns," Journal of Financial and Quantitative Analysis (September 1977), pp. 415-432.

John Affleck-Graves is an Associate Professor at the College of Business Administration, University of Notre Dame, Notre Dame, Indiana. Shantaram P. Hegde is an Associate Professor at the School of Business Administration, University of Connecticut, Storrs, Connecticut. Robert E. Miller is a Professor of Finance at the College of Business, Northern Illinois University, DeKalb, Illinois. Frank K. Reilly is the Bernard J. Hank Professor of Business Administration at the University of Notre Dame, Notre Dame, Indiana.