Docstoc

Penny Stock IPOs

Document Sample
Penny Stock IPOs Powered By Docstoc
					                                           Penny Stock IPOs
          Daniel J. Bradley, John W. Cooney, Jr., Steven D. Dolvin, Bradford D. Jordan*

We examine underpricing, long-run returns, lockup periods, and gross spreads for penny stock
IPOs over the 1990-1998 period. We find that penny stock IPOs have higher initial returns than
ordinary IPOs, but significantly worse long-run underperformance. We also find that penny
stock IPOs have longer lockup periods and larger gross spreads. To explore the effect of
potential market manipulation, we examine IPOs led by a group of underwriters that were the
subject of SEC enforcement actions and/or other penalties. Penny stock issues led by these banks
are particularly underpriced and underperform ordinary IPOs led by other underwriters.




                                   Forthcoming: Financial Management

                                                 July 15, 2005




The authors thank Ken Cyree, Tim McCormick, Jay Ritter, Ajai Singh, Lemma Senbet and James Seward (the
editors), an anonymous referee, and seminar participants at Texas Tech University and the 2002 Financial
Management Association meetings in San Antonio, TX. Dan Bradley gratefully acknowledges the financial support
of the Spiro Center for Entrepreneurial Leadership at Clemson University.

*
 Daniel J. Bradley is an assistant professor at Clemson University. John W. Cooney, Jr. is an assistant professor at
Texas Tech University. Steven D. Dolvin is an assistant professor at Butler University. Bradford D. Jordan is the
Richard W. and Janis H. Furst Endowed Chair in Finance at the University of Kentucky.
                                         Penny Stock IPOs

Introduction

   Standard research on initial public offerings (IPOs) focuses primarily on “ordinary” IPOs and

routinely uses filters to eliminate various types of offerings. Examples of groups that are almost

always purged include real estate investment trusts, unit offerings, closed-end investment funds,

and American depositary receipts. Other groups that are often dropped include spin-offs, reverse

leveraged buyouts, financial firms, mutual-to-stock conversions, and stocks with offer prices

below some minimum value. Most of these excluded subgroups have been examined

independently; however, there appears to be very little research on IPOs with low offer prices, or

so-called “penny stock” IPOs. This lack of research on low-priced IPOs is surprising, given the

prominent fraud and manipulation that existed in the 1980s that eventually led to the adoption of

the Penny Stock Reform Act of 1990 (PSRA).

   We explore the similarities and differences between penny stock IPOs and ordinary IPOs in

relation to the major findings of IPO research. In keeping with both legal standards and common

research practice, we define a penny stock IPO as an IPO that (1) is not issued by an investment

advisor (e.g., not a closed-end fund), (2) has an offer price of $5 or less, and (3) is not listed on a

national exchange or market. Thus, we classify IPOs listed on the NYSE, AMEX, or Nasdaq

National Market as ordinary IPOs regardless of the offer price. Conversely, we classify IPOs

listed on other markets (e.g., the Nasdaq SmallCap Market) as penny stock IPOs if the offering is

not issued by an investment advisor and the offer price is $5 or less.

   Given their scandal-plagued past, penny stock IPOs appear to be particularly well suited for

examining issues involving market manipulation and informational asymmetries. The firms

involved are generally small and often face limited disclosure and reporting requirements, and

they are usually underwritten by low prestige investment banks. In addition, penny stock IPOs
are much less likely to be backed by venture capital (VC), and there is likely to be little or no

institutional buying or analysis of these issues.

   Based on the 1990-1998 post-PSRA period, we find that penny stock issues show

significantly higher initial returns than do ordinary IPOs. To explore the effect of potential

market manipulation, we segregate IPOs based on whether the lead underwriter is among a group

of 39 banks that were the subject of SEC enforcement actions and/or other penalties. Penny stock

issues led by one of these 39 banks have significantly higher underpricing than do ordinary

issues led by banks not in this group.

   We next examine long-run returns. We analyze three- and five-year market-adjusted buy-

and-hold returns and find that penny stock IPOs have lower market-adjusted returns than do

ordinary IPOs. We reach the same conclusion when we examine the frequency of delistings due

to poor performance, as well as monthly abnormal performance using the Fama and French

(1993) time-series approach. We show that long-run performance for penny stock IPOs

underwritten by the 39 sanctioned banks is significantly worse than that of ordinary IPOs with

nonsanctioned underwriters and conclude that the pattern of larger initial returns and subsequent

lower long-run returns is consistent with price manipulation in the aftermarket for penny stocks

brought to the market by these sanctioned underwriters.

   In other analyses, we attempt to isolate whether the pattern of higher initial returns and

poorer long-run performance for penny stock IPOs is driven by their low offer price or by the

market or exchange on which the IPO is initially listed. Our results suggest that initial returns are

primarily related to offer price, but that both factors are related to long-run performance.

   We show that penny stock IPOs typically have much longer stated lockup periods than do

ordinary IPOs. These longer periods are consistent with the view that lockup restrictions are a

commitment mechanism to help alleviate moral hazard problems. However, we note that the lead

                                                                                                   2
underwriter can selectively allow inside shareholders of locked-up shares to sell whenever they

wish, and consequently, longer stated lockup periods might be a device to mislead investors

rather than to alleviate moral hazard problems.

   We also find that reported gross spreads for penny stock IPOs are clustered at 10% compared

to 7% for ordinary IPOs. Once we factor in additional nonaccountable expense allowances, we

find that the actual gross spreads (as opposed to the reported spreads) for penny stock IPOs are

typically 13%. Ordinary IPOs rarely have this extra component of underwriter compensation.

   The paper proceeds as follows. Section I provides background on the penny stock market.

Section II presents the data. Section III reports descriptive statistics and empirical results, and

section IV concludes.

I. Background

   The market for penny stocks has changed dramatically over the last few decades. Prior to the

development of the Nasdaq market in 1971, these stocks were typically traded over the counter

or on regional exchanges, often with very limited disclosure requirements. Particularly before the

PSRA, the penny stock market was plagued by unscrupulous broker-dealers and underwriters.

“Boiler room” tactics were common, with unregistered “brokers” cold-calling primarily

unsophisticated investors and using high pressure sales techniques to induce investment.

   In theory, a firm can undergo a reverse stock split prior to the IPO and increase its offer

price, thereby avoiding penny stock status. For example, a $10 million IPO could be packaged as

two million shares at $5 per share, or 0.2 million shares at $50 per share. However, a stockbroker

might have found it easier to convince an unsophisticated investor that it was more likely that a

stock at $5 might hit $20 than a $50 stock would go to $200. Another possible reason for

maintaining a lower offer price was that penny stocks tended to be “house stocks” with high



                                                                                                 3
percentage bid-ask spreads. A $0.50 bid-ask spread on a $5 stock might be less attention-

grabbing than a $5 bid-ask spread on a $50 stock.1

       The term “penny stock,” as it is used in the financial press and IPO research, has no precise

meaning. The definition may be based on listing requirements. For example, starting in August

1991, the Nasdaq National Market required that stocks have a minimum bid price of $5, and the

Nasdaq SmallCap Market required a minimum bid price of $3. In 1997, the minimum bid price

was changed to $4 for the Nasdaq SmallCap Market. Alternative listing criteria and discretionary

exceptions allow some firms to be listed with lower prices.

       In screening out low-priced issues, IPO researchers have used different filters. Bradley,

Jordan, and Ritter (2003) and Loughran and Ritter (2002) remove all issues that have a midpoint

of the original filing range below $8. Ljungqvist and Wilhelm (2003) and Ritter and Welch

(2002) exclude stocks with offer prices below $5, and Ritter (1991) eliminates issues with offer

prices less than $1. In their study of long-run price performance and mortality for Nasdaq-listed

issues over the period 1974-1988, Seguin and Smoller (1997) classify an issue as a penny stock

if the first listed price (as opposed to the IPO’s offer price) is less than $3. Also, many studies

drop issues that are not in the Center for Research in Security Prices (CRSP) database, which is

effectively a screen on nonlisted, lower-priced stocks.

       Although there are differences across studies, the filters commonly used in recent IPO

research closely correspond to the legal definition of a penny stock established by the PSRA. We

draw on guidelines provided by the PSRA and Section 17 (§240.3a51-1) of the Code of Federal

Regulations (CFR) and define a penny stock IPO as an issue that meets the following three

criteria: (1) it is not issued by an investment advisor (e.g., it is not a closed-end fund), (2) its

offer price is $5 or less, and (3) it is not listed on a national exchange or market. The PSRA and


1
    We thank the referee for suggesting both of these explanations.
                                                                                                  4
Section 17 of the CFR identify $5 (net of fees and commissions) as the price cutoff for

classification as a penny stock. In addition, the PSRA and Section 17 of the CFR categorize the

NYSE, AMEX, and Nasdaq National Market as a “national exchange or market.” Thus, we

classify IPOs listed on the NYSE, AMEX, or Nasdaq National Market as “ordinary” without

regard to their offer prices. IPOs listed on other markets (such as the Nasdaq SmallCap Market,

OTC Bulletin Board, pink sheets, or regional exchanges) are classified as penny stocks if the

offering is not issued by an investment advisor and the offer price is $5 or less.

   In practice, there are complications with our definition. For example, according to the CFR,

firms with sufficient net tangible assets are not penny stocks even if their offer price is less than

$5. However, lack of accounting data prevents us from considering the issuing firm’s net

tangible assets. Blind pools (blank check/shell offerings) may be classified as penny stock issues

even with offer prices above $5. Additionally, unit offerings are IPOs that include a package of a

share (or shares) of stock and warrants. Thus, a unit offering with an offer price above $5 might

also be legally classified as a penny stock if the warrant value is large enough or if multiple

shares are issued in the unit. Although many unit offerings are penny stocks, we exclude them

from our analyses because of limited data availability and accuracy. In particular, we note

substantial data errors in Thomson Financial’s SDC database with regard to units. Further

complicating matters, units frequently become unbundled once they start trading, thereby

introducing additional opportunity for data errors.

II. Data

   Our primary sample covers 1990-1998, which is the period from the implementation of the

PSRA to the beginning of the IPO “bubble” period. There are only three penny stock IPOs

during 1999-2000. Such a small sample prevents meaningful analysis, and, as we illustrate later,



                                                                                                   5
inclusion of this period distorts the comparison between ordinary and penny stock IPOs,

particularly as it relates to underpricing.

    Thomson Financial’s SDC database lists a total of 5,066 firm-commitment (ordinary and

penny stock) IPOs from 1990-1998.2 We eliminate 280 IPOs not identified on CRSP, as well as

limited partnerships (21 observations), real estate investment trusts (150 observations), closed-

end investment funds (304 observations), spin-offs (547 observations), reverse LBOs (171

observations), mutual-to-stock conversions (15 observations), American/global depositary

receipts (106 observations), units and unit trusts (361 observations), shares of beneficial interests

(one observation), and firms incorporated outside the U.S. (152 observations), leaving a sample

of 2,958 IPOs. Since CRSP only includes stocks traded on the NYSE, AMEX, Nasdaq National

Market, and Nasdaq SmallCap Market, we do not include IPOs that initially start trading on the

OTCBB, pink sheets, or regional exchanges in our dataset.

    Following our definition, we classify all offerings that initially start trading on the Nasdaq

SmallCap Market and have an offer price less than or equal to $5 as penny stock IPOs. (We do

not need to screen further for closed-end funds, since we have already eliminated these IPOs

from our sample). We define all other IPOs (i.e., those initially listed on the NYSE, AMEX, and

Nasdaq National Market regardless of offer price, and those that initially start trading on the

Nasdaq SmallCap Market with an offer price greater than $5) as ordinary IPOs. This process

results in a sample of 251 penny stock IPOs and 2,707 ordinary IPOs.

III. Descriptive Statistics and Empirical Results

    In this section, we provide descriptive statistics for our sample and the results of our

empirical analyses of penny stock and ordinary IPOs.


2
  A number of corrections were made to the SDC database, many of which were provided by Jay Ritter (see
http://bear.cba.ufl.edu/ritter/SDC%20corrections.pdf) and Alexander Ljungqvist (see Ljungqvist and Wilhelm,
2003).
                                                                                                         6
A. Descriptive Statistics

   Table I provides basic descriptive statistics on firm and offering characteristics for penny

stock and ordinary IPOs for the 1990-1998 period. We obtain the p-values from a t-test for

equality of means and assume equal or unequal variances, depending on a test of equality of

variances.

   The first line gives the average offer price for penny stock and ordinary IPOs. As indicated,

the average offer price for penny stock IPOs ($4.42) is much smaller than that of ordinary IPOs

($11.83). Offer Size (i.e., shares issued times offer price, not inflation adjusted) is smaller for

penny stock issues. In fact, 242 of the 251 penny stock IPOs have an offer size of $10 million or

less, and the maximum offer size for a penny stock IPO is $22.4 million.

                                     Insert Table I about here


   HT is a dummy variable that takes the value of one for high-tech firms, and zero otherwise.

We obtain the underlying high-tech classification codes from SDC and supplement this

information with a list of internet IPOs provided by Jay Ritter (see http://bear.cba.ufl.edu/

ritter/ipodata.htm). In our sample, approximately 32% of penny stock IPOs are high-tech firms,

compared to 42% of ordinary IPOs, which is significantly different at the 0.26% level.

   The variable VC is also a dummy variable, indicating the presence of venture capital (VC)

backing. Only 16% of penny stock IPOs are VC-backed, compared to 44% for ordinary IPOs.

These percentages are statistically different at the 0.01% level.

   The variable CM rank represents the Carter and Manaster (1990) underwriter prestige

variable as updated by Loughran and Ritter (2004). In our sample, this ranking ranges from one

(lowest prestige) to nine (highest prestige). With a mean underwriter prestige ranking of 2.6

compared to 7.1, it is clear that penny stock IPOs are primarily led by lower-ranked underwriters,

while ordinary IPOs are led by higher-ranked underwriters. We find that approximately 60% of
                                                                                                 7
underwriters that participate in the penny stock IPO market also lead ordinary IPOs. Thus, the

penny stock and ordinary IPO markets are not completely segmented.

    In Figure 1, we look more closely at the Carter and Manaster (1990) ranks. The figure shows

the percentages of ordinary and penny stock IPOs led by underwriters of each rank. Because the

underwriter ranks are ordinal, the t-test in Table 1 may not be well specified; however, Figure 1

clearly shows that there are two distinct clusters, one in the highest ranks for ordinary IPOs and

one in the lowest ranks for penny stock IPOs.

                                           Insert Figure 1 about here


    Loughran and Ritter’s (2004) ranked underwriters include a group of 39 that were subject to

enforcement actions by the SEC and/or other penalties, suspensions, and sanctions during the

1990s and early 2000s.3 For example, in 1997, the U.S. Attorney for the Southern District of

New York issued an indictment that outlined the infiltration of organized crime families into the

brokerage firm of Meyers, Pollock, & Robbins for the purpose of artificially inflating the stock

price of HealthTech International. In 2000, the Manhattan District Attorney indicted D.H. Blair

& Co., Inc., and various firm members on 173 counts for participating in the “D.H. Blair

Criminal Enterprise” during the period 1989-1998. Specific charges included stock price

manipulation and the use of illegal sales practices to generate excessive commissions. Sterling

Foster & Company, Inc., was charged in 1997 with using boiler room sales tactics to illegally

obtain $75 million from investors in connection with the IPOs of six companies. We create a

dummy variable, Enforce, that is equal to one if the lead underwriter is in this group of 39. Table

I shows that potentially manipulative activity is more prevalent in penny stock IPOs.



3
 We thank Jay Ritter for supplying us with an initial list of 32 underwriters. The original source for this information
on enforcement actions was the Chicago office of the SEC’s Division of Enforcement for 1995-1999. Additional
searching produced the names of another seven underwriters.
                                                                                                                     8
Approximately 26% of these IPOs are led by this set of underwriters compared to only 4% for

ordinary IPOs.

    The remaining variables in Table I are those that have been most closely examined in the IPO

literature: initial return, lockup length, gross underwriting spread, and long-run return. As the

table shows, penny stocks have higher initial returns (22.4% compared to 15.4%), longer lockup

periods (452 days compared to 208 days), and larger reported gross spreads (9.7% compared to

7.2%). In each case, the difference is statistically significant and economically meaningful.

    To examine long-run performance, we compute the total dividend- and split-adjusted buy-

and-hold percentage return beginning at the end of the first trading day. By doing so, we exclude

the initial return. The ending date is either 756 trading days (i.e., three years) after the first day of

trading or the delisting return date, whichever is earlier. We use the CRSP delisting return

whenever available. When it is not available, we assume that the delisting return for

performance-related delists is -30% for NYSE/AMEX stocks, -55% for Nasdaq stocks, and zero

for non-performance delists (Shumway, 1997; Shumway and Warther, 1999). We label this

variable 3YrRet. Penny stock IPOs have a lower mean 3YrRet than do ordinary IPOs, -21.7%

compared to 44.4%. The difference in means between the two groups is statistically significant at

the 0.01% level. We then adjust 3YrRet by subtracting the contemporaneous compounded return

on the CRSP Nasdaq value-weighted market index and label this variable 3YrMARet. Both penny

stock IPOs and ordinary IPOs have negative market adjusted returns (-101.8% compared to -

39.9%), which is consistent with results reported in previous studies (e.g., Ritter, 1991). The

difference is statistically significant at the 0.01% level.

B. Initial Returns

    Figure 2 plots the average initial returns by year for penny stock and ordinary IPOs. From

1990 to 1998, penny stock IPOs are more underpriced in every year except 1998. The 1998 result

                                                                                                       9
may be due in part to the beginning of the IPO “bubble” period, during which there were very

few penny stock IPOs and, at the same time, extraordinarily large initial returns for ordinary

IPOs. Of the 252 offerings in 1998, only 14 (or 5.6%) were penny stock IPOs, the lowest

percentage across the eight years of our study. In fact, if we include the 1999-2000 period in our

sample, ordinary IPOs are more underpriced, on average, than penny stock IPOs (26.2%

compared to 22.2%). However, this result is misleading, because, as noted earlier, there are only

three penny stock IPOs during the 1999-2000 period.

                                      Insert Figure 2 about here


    The significant difference in initial returns between ordinary and penny stock IPOs may be

due to factors commonly discussed in the IPO literature, such as adverse selection and

informational asymmetry. These factors may be particularly pronounced in penny stock

offerings. However, since sanctioned underwriters lead a substantial percentage of the penny

stock offerings in our sample, it is also possible that manipulation drives up first-day market

prices beyond their equilibrium levels, leading to higher initial returns. As a first look at this

issue, we segment both penny stock IPOs and ordinary IPOs based on the variable Enforce.

Table II reports the results of this analysis.

                                      Insert Table II about here


    The results in Table II show that penny stock IPOs led by the sanctioned investment banks

have a mean initial return of 31.6% compared to 19.2% for issues not led by this group. These

means are significantly different at the 5% level. For ordinary IPOs, there is no statistical

difference between the mean initial return for issues led by the 39 sanctioned underwriters

(13.3%) and other issues (15.5%). However, the mean initial return for the sample of penny

stock IPOs with sanctioned underwriters (31.6%) is significantly higher at the 1% level than is

                                                                                               10
the mean for ordinary IPOs with other underwriters (15.5%). These results suggest that a large

portion of the difference in initial returns between the two types of IPOs could be due to

manipulation.

    To examine the underpricing for penny stock and ordinary IPOs in more detail, we estimate

the following regression, which controls for important determinants of underpricing identified by

previous studies:

    Initial return = α + β1PS + β2Enforce + β3PS × Enforce + β4VC + β5VC × Enforce
        + β6Over + β7HT + β8MedRank + β9HighRank + β10Secondary + β11LnSize
        + β12Integer + β13IPOLag+ β14NasLag + β15Multiple + β16PartU
        + β17PartD + β18-25Year Dummies + ε                                                       (1)

where

Initial return      = percentage return from the offer price to the closing price on the first day of
                        trading
PS                  = dummy variable equal to one if the issue is a penny stock, zero otherwise
Enforce             = dummy variable equal to one if the lead underwriter is one of a group of 39
                        that were the subject of SEC enforcement actions and/or other penalties
                        during the 1990s or early 2000s, zero otherwise
PS × Enforce        = interaction of PS and Enforce
VC                  = dummy variable equal to one if the issue is VC-backed, zero otherwise
VC × Enforce        = interaction of VC and Enforce
Over                = “overhang,” measured as pre-IPO shares retained divided by total shares
                        offered
HT                  = dummy variable equal to one if the issuing firm is high-tech, zero otherwise
MedRank             = dummy variable equal to one if the lead underwriter has a Carter and
                        Manaster (1990) rank greater than or equal to six and less than eight, zero
                        otherwise
HighRank            = dummy variable equal to one if the lead underwriter has a Carter and
                        Manaster (1990) rank greater than or equal to eight, zero otherwise
Secondary           = percentage of shares offered that are secondary shares
LnSize              = natural logarithm of the CPI-adjusted (1982-1984=100) offer size in
                        millions of dollars
Integer             = dummy variable equal to one if the offer price is an integer, zero otherwise
IPOLag              = an underpricing index defined as the average initial return, in percent, for all
                        IPOs in the sample for the month before the issue date
NasLag              = compounded return, in percent, for the CRSP Nasdaq value-weighted
                        market index for the 21 trading days (i.e., one month) prior to the issue
                        date
                                                                                                   11
Multiple           = dummy variable equal to one if the firm has more than one class of common
                      shares, zero otherwise
PartU              = percentage difference between the offer price and the original mid-file price
                      if the adjustment is positive, zero otherwise
PartD              = percentage difference between the offer price and the original mid-file price
                      if the adjustment is negative, zero otherwise, and,
Year Dummies       = eight dummy variables equal to one if the offering is in year 1991, … ,
                      1998, zero otherwise

   The first two variables, PS and Enforce, are dummy variables for penny stock status and SEC

enforcement actions and/or other penalties, respectively. We include a dummy variable

indicating venture capital backing (VC), which we also interact with Enforce (VC × Enforce). The

remaining variables are among those widely used in various recent studies of IPO underpricing

(e.g., Bradley and Jordan, 2002; Cai, Ramchand, and Warga, 2004; Houge, Loughran, Suchanek,

and Yan, 2001; Loughran and Ritter, 2004).

   Issues with greater overhang (Over) are typically more underpriced over our sample period,

as are firms in high-tech industries (HT). Contrary to earlier findings (e.g., Carter and Manaster,

1990), numerous recent studies find that firms associated with prestigious lead underwriters are

more underpriced in the 1990s (e.g., Beatty and Welch, 1996; Cooney, Singh, Carter, and Dark,

2001; Dolvin, 2005). Therefore, we include dummy variables for high-prestige banks (ranks

eight and higher, HighRank) and medium-prestige banks (ranks greater than or equal to six, but

less than eight, MedRank). The omitted category of low-prestige banks includes all of the

sanctioned underwriters, most of which are ranked two or lower.

   Following Habib and Ljungqvist (2001), we include the percentage of secondary shares sold

in the IPO (Secondary). Habib and Ljungqvist (2001) find a negative relation between initial

returns and this variable. The natural log of the CPI-adjusted size of the deal (LnSize) is a

common conditioning variable, but it also controls for the possibility that the relation between

penny stock status and initial returns is simply a size effect. A recent study by Bradley, Cooney,

Jordan, and Singh (2004) finds that IPOs priced on an integer (Integer) exhibit greater
                                                                                    12
underpricing. The lag variables, IPOLag and NasLag, proxy for the existence of a hot IPO

market and investor sentiment.

    We include Multiple to follow Smart and Zutter (2003), who find that IPOs with multiple

share classes are less underpriced. The listing of firms with multiple classes of stock comes from

Smart and Zutter (2003), as tracked by Jay Ritter. Because this listing of firms does not include

low-priced issues, there may be some penny stock IPOs that have more than one class of

common stock, but are not identified as such in our sample. To avoid potential bias associated

with this variable, we drop Multiple and repeat the analysis in columns (5) through (8) in Table

III, finding results similar to those reported.

    Finally, PartU and PartD control for the well-known partial adjustment phenomenon

documented by Hanley (1993). Following Bradley and Jordan (2002), we include separate

variables for upward and downward adjustments because recent studies find evidence of an

asymmetric effect (upward adjustments have a much greater impact). Our results are presented in

Table III. Throughout this study, p-values from ordinary OLS t-statistics are reported.

                                      Insert Table III about here


    Regressions (1) through (5) in Table III examine all IPOs in our sample. In regression (1), the

only independent variable is the dummy PS. Consistent with Table I, the results show that penny

stock IPOs are more underpriced than ordinary IPOs by 7%, which is significant at any

conventional level. Regression (2) adds the variable Enforce. The penny stock dummy, PS,

remains economically and statistically significant, but Enforce is nonsignificant.

    Our previous results in Table II suggest that penny stock IPOs led by sanctioned underwriters

are particularly underpriced; therefore, in regression (3) we include an interaction between PS

and Enforce. Consistent with our earlier findings, PS remains significant (p = 0.0549) and

Enforce remains nonsignificant. However, the interaction between PS and Enforce is relatively
                                                                                                13
large (14.6%) and highly significant (p = 0.0009). Thus, the evidence so far indicates that penny

stocks are more underpriced than are ordinary IPOs, with particularly severe underpricing for

penny stock issues led by underwriters that were the subject of SEC enforcement actions and/or

other penalties.

    To examine the effect of VC participation, we include two additional variables in regression

(4), the dummy VC and the interaction between VC and Enforce. VC backing is positively

related to underpricing, but subsequent analysis shows that this result may be due to industry

proxy effects. Of greater interest here is that the coefficient on the interaction term, -19.5%, is

economically and statistically significant. This finding indicates that the presence of a venture

capitalist substantially reduces the underpricing associated with the group of sanctioned

underwriters. A plausible explanation is that the presence of a venture capitalist, who has

reputational capital at stake, reduces the likelihood of manipulation, and so plays a valuable

certification role in offers underwritten by these particularly low-prestige banks.4

    Regression (5) adds the other variables listed in Equation (1) above. The coefficient on PS is

6 and is highly significant. The interaction between PS and Enforce also remains relatively large

at 10.7 and significant. The VC dummy is no longer significant, but the interaction between VC

and Enforce remains negative and significant with a coefficient of -11. The remaining variables

in regression (5) have signs and significance levels that are generally consistent with previous

studies.

    In comparing regressions (4) and (5), the VC dummy moves from positive and highly

significant to nonsignificant. As discussed in Bradley and Jordan (2002), the reason VC-backed

issues appear to be more underpriced is probably due to industry focus—VCs tend to concentrate

4
  An interesting issue arises on the interrelation between VC backing and the group of sanctioned underwriters. The
referee points out that VC firms listed in a manipulated offering might not be genuine. To explore this possibility,
we examined prospectuses from EDGAR on a small sample of penny stock IPOs with sanctioned underwriters and
VC backing (as reported by SDC). In each case, the VC firm listed in the IPO prospectus appeared to be a legitimate
VC firm based on industry sources and web sites.
                                                                                                                 14
in industries with higher underpricing. In our case, for example, adding only the high tech

dummy variable, HT, to regression (4) eliminates any significant effect from the VC dummy.

    In unreported results, we repeat regression (5) of Table III after dropping all IPOs with initial

mid-file prices at or above $10.5 Our motivation for this analysis is the possibility that lower-

priced IPOs are simply more underpriced than are higher priced IPOs, and that penny stock

status per se is not important. Higher initial returns for penny stock IPOs are consistent with

Fernando, Krishnamurthy, and Spindt (2004), who find a large degree of underpricing for IPOs

with low offer prices. However, Fernando et al. do not specifically examine penny stock status.

We find that PS drops in level (3.5) and significance (p = 0.1256), and the coefficient on Enforce

remains nonsignificant. However, the coefficient on the interaction of PS and Enforce becomes

larger (13.1) and more significant (p = 0.0013). An interesting result from the lower-priced

sample is that the coefficient on VC is negative (-4.6) and significant (p = 0.0191), which

suggests that VC certification is more important for lower-priced issues. Also, the coefficient on

the VC × Enforce interaction remains significantly negative (coefficient = -8.1, p = 0.0977).

    An important aspect of our analysis of the lower-priced sample is our use of the original mid-

file price as the cutoff. If we were to use the offer price, we would create a potentially serious

bias because of the partial adjustment phenomenon. Firms with offer prices originally below

(above) $10, but subsequently adjusted to more (less) than $10, would be excluded (included).

The resulting sample would have a downward bias to its measured underpricing because firms

with downward adjustments would be overrepresented relative to firms with upward

adjustments. The $5 offer price cutoff we use for defining penny stock status creates a similar,

5
  When a firm files its preliminary prospectus with the SEC, an initial file range is given that represents a range of
possible offer prices for the issuing firm’s stock. During the registration process, the IPO firm and its underwriters
will typically go on a road show pitching the IPO to institutional investors. Using the information gleaned from the
road show, and other sources, the offer price is set, which may be within, above, or below the initial file range. For
penny stock IPOs, however, it is unlikely that a road show will be undertaken because institutional investors would
not be interested in such small offerings (or might be prohibited from purchasing them). Thus, the underwriter may
need to set the offer price for penny stock IPOs without the benefit of institutional input.
                                                                                                                  15
but unavoidable, problem. However, in this case, the effect is that we are biased against finding

greater underpricing for penny stock IPOs. That is, our estimates of the difference in

underpricing for penny stock IPOs may be downward biased. In fact, during the 1990-1998

period, there are 25 penny stock IPOs (less than 1% of our total IPO sample) that meet this

criterion (i.e., have an original mid-file price above $5, but a final offer price of $5 or less), and

the average initial return for these 25 IPOs is only 11.7%. As noted earlier, the average initial

return for our entire sample of penny stock IPOs is 22.4%.

   We note that because of changes in issue characteristics and other aspects of the IPO market,

there may be some confounding of time-series and cross-sectional effects in our regressions.

Thus, we split the sample into three subperiods, IPOs offered during 1990-1992, 1993-1995, and

1996-1998, and repeat regression (5) in Table III for each subperiod. Our main variable, PS,

remains positive in all three subperiods, with p-values of 0.0005, 0.1574, and 0.2255,

respectively. The two other variables of interest, PS×Enforce and VC×Enforce, are both

nonsignificant in the 1990-1992 period, but there are very few IPOs led by sanctioned

underwriters during this period. Similar to the results in regression (5) of Table III, PS×Enforce

is positive and significant in the later two periods with p-values of 0.0635 and 0.0172,

respectively. The interaction of VC and Enforce is negative and significant (p-value = 0.0109) in

the 1993-1995 period. For 1996-1998, VC×Enforce is also negative, but with a p-value of

0.1949.

   It is possible that the control variables used in our Table III regressions affect ordinary IPOs

and penny stock IPOs differently. Therefore, we estimate regression (5) from Table III separately

for ordinary and penny stock IPOs. We note that in these analyses, we must drop the dummy

variable PS (and the interaction of PS and Enforce). The results from these analyses are

presented in regressions (6) and (7), respectively, of Table III. Column (8) of Table III reports

                                                                                                   16
the p-value for difference tests between the marginal effects (i.e., the estimated coefficients) in

each of these regressions.

   The findings of the ordinary IPO analysis in regression (6) are generally consistent with those

from the entire sample. One result of particular interest is that the interaction of VC and Enforce

remains negative even with the penny stocks excluded, but the coefficient is significant at only

the 16% level. For the penny stock IPOs, only the variables Enforce, the interaction of VC and

Enforce, Over, and IPOLag are significant (at the 10% level) in predicting initial returns. The

lack of significance for other variables may be due to the relatively small sample size. In many

cases, the coefficients in the regressions (6) and (7) are numerically similar. In fact, as column

(8) shows, the hypothesis that the coefficients are identical can be rejected at the 10%

significance level for only six variables: Enforce, Over, IPOLag, Multiple, PartU, and the 1998

year dummy (not reported).

   Lee and Wahal (2004) conclude that VC backing is an endogenous variable. Thus, as a final

robustness check, we repeat columns (4) through (8) in Table III, using a two-stage least squares

framework. The results for PS, Enforce, and PS×Enforce, are similar to those presented in Table

III. For Predicted VC, the results are similar to Table III for regression (4) through (6), but the

coefficient is now negative and significant in regression (7). Predicted VC×Enforce, however, is

nonsignificant in all regressions. Consistent with this last result, we find that none of the

predicted venture-backed penny stock IPOs from the first stage regression are brought to the

market by a sanctioned underwriter.

C. Long-Run Returns

   We next compare long-run returns for ordinary and penny stock IPOs. If the greater

underpricing for penny stock IPOs is driven primarily by economic forces such as informational

asymmetry, then the pricing effects should be incorporated into the initial offer and market

                                                                                                17
prices. Therefore, we would expect similar long-run performance for penny stock and ordinary

IPOs. However, if price manipulation drives higher penny stock underpricing, then we would

expect penny stock IPOs to show lower long-run performance as the post-IPO prices of

manipulated issues revert to fundamental levels.

   As noted in Table I, the sample of penny stock IPOs has a significantly lower mean three-

year market-adjusted buy-and-hold return than the sample of ordinary IPOs. We also calculate

the market-adjusted buy-and-hold return for the five years following the IPO date for penny

stock and ordinary IPOs, 5YrMARet, and present these results in the second row of Table IV.

Similar to the three-year results, the mean market-adjusted return calculated over a five-year

period for penny stock IPOs is significantly lower than the mean for ordinary IPOs, -126.4% and

-55.2%, respectively.

                                    Insert Table IV about here


   In Table IV, we further explore long-term performance. We first examine the percentage of

IPOs delisted by the NYSE, AMEX, or Nasdaq because of liquidation or poor performance

within three years (and five years) of the IPO date (rows three and four). We identify liquidations

as firms with CRSP delist codes of 400-490; performance-related delists are codes 500 and 520-

591. Over the three-year period following the IPO, penny stock IPOs have a significantly higher

percentage of firms that delist for liquidation or performance reasons (31.5%) than do ordinary

IPOs (6.4%). The most frequent reasons given for delisting are insufficient float, capital, or

assets, low stock price, delinquencies in filing, or nonpayment of fees. Analysis over a five-year

period shows similar results.

   In a related study, Beatty and Kadiyala (2003) also find that penny stock IPOs underperform

relative to ordinary IPOs and are more likely to delist for performance-related reasons. They

investigate whether the PSRA curtailed fraudulent securities issues by comparing IPOs issued
                                                                                                18
during the pre-PSRA period to those issued during the post-PSRA period. Their results indicate

that it did not. In particular, they find that measures of issuer quality for non-penny stock IPOs

declined significantly between the two periods, suggesting that issuers dodged the new rules by

simply issuing securities at higher prices.

    We next evaluate long-run performance using the Fama and French (1993) time-series

approach. To do so, we form monthly value-weighted portfolios for each calendar month. These

portfolios comprise all IPOs issued during the previous 60 calendar months (not including the

month of issue) that are still CRSP-listed. If the firm delists before the end of the five-year

period, we include the delisting return as the final month of the return series. We then regress the

monthly value-weighted portfolio returns less the risk-free rate (i.e., NetReturn) on the returns

from the three Fama and French (1993) factor portfolios Mkt, SMB, and HML. Mkt is the excess

value-weighted total market return, SMB is the return from a portfolio of small capitalization less

big capitalization stocks, and HML is the return from a portfolio of high minus low book-to-

market firms.6 To reduce the influence of an individual stock on the results, we eliminate a

particular month’s observation if there are fewer than ten stocks in the portfolio. We find similar

results when cutoffs of five or 15 stocks are used.7 Equation (2), estimated over the period from

January 1990 to December 2003, or 168 calendar months, is as follows:

    NetReturn = α + β1Mkt + β2SMB + β3HML + ε                                                                   (2)

    In Equation (2), the intercept (alpha) represents the monthly abnormal return. We use only

139 of the 168 calendar months in the penny stock regressions and 164 of 168 months for the

ordinary IPO regressions. The excluded months do not have at least ten stocks with return data.




6
  The factor portfolio returns are from Ken French’s website (mba.tuck.dartmouth.edu/pages/faculty/ken.french),
where greater detail about their construction is available.
7
  As an alternative, we include all months, but use t-statistics constructed with White’s (1980) heteroskedasticity-
consistent standard errors. Results are similar to those presented in Table IV.
                                                                                                                19
Consistent with our earlier results, we find that penny stock IPOs have an alpha equal to -1.69%,

significant at the 1% level, but the alpha for ordinary IPOs is close to zero.

   To test for differences in abnormal returns, we form a zero-cost portfolio that is long in

ordinary IPOs and short in penny stocks IPOs and re-estimate Equation (2), so that the intercept

is the difference in abnormal returns. To be included in the difference regressions, the calendar

month must include at least ten ordinary IPO stock returns and ten penny stock IPO stock

returns. For penny stock and ordinary IPOs, the difference in abnormal performance of 1.71%

per month is statistically significant with a p-value of 0.0136 (N = 139 calendar months).

   To further evaluate the effects of potential manipulation, we again split the sample into two

groups based on the variable Enforce, and compare penny stock IPOs led by sanctioned

underwriters to ordinary IPOs led by other underwriters. Rows 6-20 of Table IV present our

results. Consistent with the presence of market manipulation, according to all three long-run

performance measures, the sample of penny stock IPOs with sanctioned underwriters

underperforms the sample of ordinary IPOs with nonsanctioned underwriters. The differences in

performance measures are quite large. For example, the mean three-year, market-adjusted, buy-

and-hold return is -131.8% for penny stock/sanctioned IPOs compared to -37.3% for

ordinary/nonsanctioned IPOs. The frequency of performance delists over the three years

following the IPO is more than four times larger for penny stock/sanctioned IPOs as compared to

ordinary/nonsanctioned offerings. The Fama and French (1993) alpha for penny stock/sanctioned

IPOs indicates abnormal performance of -2.73% per month, and the alpha for

ordinary/nonsanctioned issues is 0.02%.

   Further examination of Table IV shows that long-run underperformance is primarily a result

of penny stock status. Compared to penny stock IPOs led by a sanctioned underwriter, penny

stock IPOs led by nonsanctioned underwriters have higher three- and five-year market-adjusted

                                                                                              20
buy-and-hold returns. However, the frequency of performance delists and Fama and French

(1993) alphas are statistically similar for these two groups. It is interesting to note that for

ordinary IPOs, the evidence that being led by a sanctioned underwriter results in poorer long-

term performance is much stronger.

   The analysis of long-run returns for sanctioned compared to nonsanctioned underwriters

presents a potential endogeneity problem. That is, it is unlikely that an underwriter would have

been sanctioned unless at least one of their IPOs had significantly declined in price. Thus, the

differential performance between these two groups may be overstated.

D. Offer Price Compared to Exchange/Market

   All 251 penny stock IPOs in our study are listed on the Nasdaq SmallCap Market. However,

our sample of ordinary IPOs includes both high-priced offerings that start trading on the Nasdaq

SmallCap Market and low-priced offerings that start trading on the AMEX or Nasdaq National

Market. To explore the differential impact of high compared to low offer prices and the

particular exchange or market on which the IPO starts trading, Table V provides mean initial

returns, three- and five-year market-adjusted returns, and three- and five-year delisting

frequencies classified by whether the IPO has a high or low offer price (i.e., greater than $5

compared to less than or equal to $5) and the exchange or market where the IPO initially starts

trading. We note that the limited sample sizes prevent a meaningful analysis of long-run returns

using the Fama and French (1993) time-series approach.

                                     Insert Table V about here


   Table V shows that there are no IPOs with offer prices less than or equal to $5 listed on the

NYSE and only nine on the AMEX. However, we find larger numbers of low-priced offerings

for IPOs listed on the Nasdaq National Market (48) and Nasdaq SmallCap Market (251).


                                                                                             21
   When we focus on the Nasdaq National and SmallCap markets (because of their larger

sample sizes), our examination of the point estimates shows that initial returns are higher and

long-run performance is poorer for low-priced stocks across both markets. In addition, holding

the offer price range constant, IPOs listed on the Nasdaq SmallCap Market have similar initial

returns, but lower long-run performance than do the IPOs listed on the Nasdaq National Market.

The group with the highest initial returns and lowest long-run performance is our sample of 251

penny stocks. Several of the pair-wise comparisons do not exhibit statistically different means.

However, this lack of significance might be driven by the relatively small sample sizes.

   Overall, our results suggest that differences in initial returns between penny stock and

ordinary IPOs are primarily related to offer price, but both the offer price and the stock market

where the IPO starts trading are related to long-run performance.

E. Lockup Lengths

   Almost all IPOs are subject to a lockup agreement. This agreement essentially restricts

insiders from selling shares for a prespecified time after the IPO without the written consent of

the lead underwriter. Bradley, Jordan, Roten, and Yi (2001) and Field and Hanka (2001) find that

the length of the lockup period for ordinary IPOs has become standardized at 180 days.

   Table VI contains an analysis of lockup lengths for penny stock IPOs, broken out by VC- and

non-VC-backed issues. Because Bradley, Jordan, Roten, and Yi (2001) find that reported lockup

periods of zero days in the SDC database are generally data errors, we drop issues with reported

lockup periods of zero in all analyses related to lockup periods.

                                    Insert Table VI about here


   The conventional wisdom is that lockups serve as a commitment device designed to alleviate

moral hazard and asymmetric information problems. Longer lockup times may be required when


                                                                                              22
such asymmetry, or the potential for fraud, are especially pronounced. Consistent with this idea,

Table I shows that penny stock IPOs have longer lockup periods, on average, than do ordinary

IPOs (452 days compared to 208 days, respectively). Table VI shows that 19% of penny stock

IPOs have lockups of 180 days or less. In comparison, 86% of ordinary IPOs have lockups of

180 days or less (not reported). VC-backed firms in the penny stock IPO market have an average

lockup period of 359 days, compared to 473 days for non-VC-backed firms, which is statistically

different at the 1% level. This finding supports the proposition that VC backing reduces agency

problems in the penny stock IPO market.

   In contrast to our results on initial and long-run returns, when we segregate our sample based

on Enforce, we find no effect on lockup lengths for penny stock IPOs. For ordinary IPOs, issues

led by the sanctioned underwriters have significantly longer lockup lengths. In this case, the

difference probably reflects the general tendency for lockup lengths to be shortest for larger

issues led by higher-prestige underwriters.

   Caution in warranted in interpreting differences in stated lockup lengths, particularly when

fraud and/or manipulation are a concern. Since insiders can be released from lockups at the

underwriter’s discretion, longer stated lockup periods might be used to fool investors into

thinking that insiders will not sell out before the end of the stated period, when in fact the

underwriter plans to release some (or all) of the shares early. Because public disclosure of early

release is not required, it is difficult to determine if underwriters are playing this game.

F. Gross Spreads

   Referring again to Table I, average reported gross spreads are higher for penny stock IPOs

(9.7%) than for ordinary IPOs (7.2%). Higher gross spreads for penny stock IPOs are consistent

with economies of scale in the offer process, as well as with the greater expenditures (per dollar

raised) needed to value and market these more informationally problematic issues.

                                                                                               23
    Chen and Ritter (2000) find that over 90% of moderate-sized equity IPOs (i.e., $20 million to

$80 million in gross domestic proceeds, expressed in 1997 dollars) from 1995 to 1998 have gross

spreads equal to 7%. Over the period we study, penny stock IPOs have a similar clustering

around 10%. In fact, over the 1990-1998 period, 80.8% of penny stock issues have reported

gross spreads of 10% while 74.9% of ordinary IPOs have gross spreads of 7%.8 Thus, we find

that a single fixed spread does not characterize the IPO market. Rather, IPO spreads tend to

cluster around two amounts: 7% for ordinary IPOs and 10% for penny stock IPOs.

    In unreported results, we also examine spreads from the 1980s. For penny stock IPOs, 81.3%

have a reported gross spread equal to 10% during this period. Thus, it appears that penny stock

gross spreads have remained relatively constant through time. In contrast, as Chen and Ritter

(2000) note, the clustering of moderate-sized IPO spreads around 7% is a more recent

phenomenon.

    The difference in gross spreads between penny stock and ordinary IPOs appears to be even

wider than reported in Table I. When we reviewed prospectuses for penny stock IPOs listed on

the SEC’s EDGAR database for the years 1997 and 1998, we found that for all penny stock IPOs

with a 10% direct gross spread, the footnote to the underwriter’s commission included a

nonaccountable allowance, almost always equal to 3%. We also examined a sample of ordinary

IPOs over the same period, none of which include this allowance. So, penny stock IPOs typically

have a true spread of 13%.

    Barry, Muscarella, and Vetsuypens (1991) observe that breaking the spread into two pieces

may allow the underwriter to avoid certain NASD restrictions and state “blue-sky” laws

regarding maximum permitted spreads. Second, Barry et al. note that warrants are prevalent in



8
 To account for small rounding differences, IPOs with per share gross spreads within 1 penny of 10% (7%) of the
offer price are defined as having 10% spreads (7% spreads).
                                                                                                            24
smaller, more informationally problematic issues. Thus, the use of underwriter warrants may

provide even more compensation to underwriters of penny stock IPOs.

   After segregating based on Enforce, we find no difference in reported gross spreads for the

penny stock IPOs. Ordinary IPOs led by sanctioned underwriters have significantly greater

reported spreads than issues led by other underwriters. However, as with lockup lengths, this

difference is probably due to issue size. There is a general tendency for smaller issues led by

lower prestige underwriters to have larger spreads.

IV. Conclusions

   IPO researchers routinely screen out lower-priced issues and, as a consequence, penny stock

IPOs have been largely ignored in the IPO literature. For the post-Penny Stock Reform Act

(PSRA) era, after we control for a wide range of issue and market characteristics, we find that

penny stock IPOs are significantly more underpriced than are ordinary IPOs. When we examine

long-run returns, we find that penny stock IPOs underperform ordinary IPOs. Further tests

suggest that higher initial returns for penny stock IPOs are primarily related to their low offer

price. However, both the IPO’s offer price and the market or exchange on which it initially starts

trading are related to the IPO’s long-run performance.

   We separately examine IPOs led by a group of 39 underwriters that were the subject of

various penalties and enforcement actions in the 1990s and early 2000s. These underwriters

account for 26% of our penny stock sample and 4% of our ordinary IPOs. The most notable

result is that penny stock IPOs led by these 39 underwriters are significantly more underpriced

and subsequently perform significantly worse over the long run than do ordinary IPOs led by

other underwriters. This pattern of high initial returns, followed by especially poor long-run

performance for penny stock issues led by these underwriters, suggests that market manipulation

and/or fraud may explain a substantial part of the underpricing and long-run underperformance

                                                                                               25
of penny stock IPOs. These results also suggest that despite reforms such as the PSRA, the penny

stock market remains a potentially fertile ground for market manipulation and fraud.

   We provide evidence that penny stock IPOs have substantially longer lockup periods and

larger gross spreads than do ordinary IPOs. Longer lockup periods for penny stock IPOs can

serve as a commitment device to reduce problems associated with moral hazard and asymmetric

information. However, because underwriters can release shares from the lockup at their

discretion without public notification, it is not clear how effective this device would be for penny

stock IPOs if the underwriters are involved in stock price manipulation. We find that there are

shorter lockups lengths for penny stock IPOs with venture capital backing, which suggests a

potential certification role for VC firms. Higher gross spreads for penny stock IPOs are

consistent with economies of scale in the offer process, as well as with the greater expenditures

(per dollar raised) needed to value and market these more informationally problematic issues.




                                                                                                 26
                                          References

Barry, C., Muscarella, C., and M. Vetsuypens, 1991, “Underwriter Warrants, Underwriter
Compensation, and the Costs of Going Public,” Journal of Financial Economics 29, 113-135.

Beatty, R. and P. Kadiyala, 2003, “Impact of the Penny Stock Reform Act of 1990 on the IPO
Market,” Journal of Law and Economics 46, 517-542.

Beatty, R. and I. Welch, 1996, “Issuer Expenses and Legal Liability in Initial Public Offerings,”
Journal of Law and Economics 39, 545-602.

Bradley, D., Cooney, J., Jordan, B., and A. Singh, 2004, “Negotiation and the IPO Offer Price: A
Comparison of Integer Versus Non-integer IPOs,” Journal of Financial and Quantitative
Analysis 39, 517-540.

Bradley, D. and B. Jordan, 2002, “Partial Adjustment to Public Information and IPO
Underpricing,” Journal of Financial and Quantitative Analysis 37, 595-616.

Bradley, D., Jordan, B., and J. Ritter, 2003, “The Quiet Period Goes out with a Bang,” Journal of
Finance 58, 1-36.

Bradley, D., Jordan, B., Roten, I., and H. Yi, 2001, “Venture Capital and IPO Lockup
Expiration: An Empirical Analysis,” Journal of Financial Research 24, 465-493.

Cai, N., Ramchand, L., and A. Warga, 2004, “The Pricing of Equity IPOs that Follow Public
Debt Offerings,” Financial Management 33, 5-26.

Carter, R. and S. Manaster, 1990, “Initial Public Offerings and Underwriter Reputation,” Journal
of Finance 45, 1045-1067.

Chen, H. and J. Ritter, 2000, “The Seven Percent Solution,” Journal of Finance 55, 1105-1131.

Cooney, J., Singh, A., Carter, R., and F. Dark, 2001, “IPO Initial Returns and Underwriter
Reputation: Has the Inverse Relationship Flipped in the 1990s?” Texas Tech University, Case
Western Reserve University, and Iowa State University Working Paper.

Dolvin, S., 2005, “Market Structure, Changing Incentives, and Underwriter Certification,”
Journal of Financial Research 28 (Forthcoming).

Fama, E. and K. French, 1993, “Common Risk Factors in the Returns on Stocks and Bonds,”
Journal of Financial Economics 33, 3-56.

Fernando, C., Krishnamurthy, S., and P. Spindt, 2004, “Are Share Price Levels Informative?
Evidence from the Ownership, Pricing, Turnover and Performance of IPO Firms,” Journal of
Financial Markets 7, 377-403.

Field, L. and G. Hanka, 2001, “The Expiration of IPO Share Lockups,” Journal of Finance 56,
471-500.
                                                                                              27
Habib, M. and A. Ljungqvist, 2001, “Underpricing and Entrepreneurial Wealth Losses in IPOs:
Theory and Evidence,” Review of Financial Studies 14, 433-458.

Hanley, K., 1993, “The Underwriting of Initial Public Offerings and the Partial Adjustment
Phenomenon,” Journal of Financial Economics 34, 231-250.

Houge, T., Loughran, T., Suchanek, G., and X. Yan, 2001, “Divergence of Opinion, Uncertainty,
and the Quality of Initial Public Offerings,” Financial Management 30, 5-23.

Lee, P. and S. Wahal, 2004, “Grandstanding, Certification and the Underpricing of Venture
Capital Backed IPOs,” Journal of Financial Economics 73, 375-407.

Ljungqvist, A. and W. Wilhelm, 2003, “IPO Pricing in the Dot-com Bubble,” Journal of Finance
58, 723-752.

Loughran, T. and J. Ritter, 2002, “Why Don’t Issuers Get Upset About Leaving Money on the
Table in IPOs?” Review of Financial Studies 15, 413-443.

Loughran, T. and J. Ritter, 2004, “Why Has IPO Underpricing Changed over Time?” Financial
Management 33, 5-37.

New York State Attorney General, 1997, “Report on Microcap Stock Fraud,” Bureau of Investor
Protection and Services.

Ritter, J., 1991, “The Long-run Performance of Initial Public Offerings,” Journal of Finance 46,
3-27.

Ritter, J. and I. Welch, 2002, “A Review of IPO Activity, Pricing, and Allocations,” Journal of
Finance 57, 1795-1828.

Seguin, P., and M. Smoller, 1997, “Share Price and Mortality: An Empirical Evaluation of
Newly Listed Nasdaq Stocks,” Journal of Financial Economics 45, 333-363.

Shumway, T., 1997, “The Delisting Bias in CRSP Data,” The Journal of Finance 52, 327-340.

Shumway, T. and V. Warther, 1999, “The Delisting Bias in CRSP’s Nasdaq Data and Its
Implications for the Size Effect,” Journal of Finance 54, 2361-2379.

Smart, S. and C. Zutter, 2003, “Control as a Motivation for Underpricing: A Comparison of
Dual- and Single-class IPOs,” Journal of Financial Economics 69, 85-110.

White, H., 1980, “A Heteroskedasticity-Consistent Covariance Matrix and a Direct Test for
Heteroskedasticity,” Econometrica 48, 817-838.




                                                                                             28
                                       Table I. Firm Characteristics

This table reports means and standard deviations for penny stock and ordinary IPOs. We define penny
stock IPOs as offerings that are not issued by an investment advisor (e.g., not a closed-end fund), have an
offer price at or below $5, and are not listed on a national exchange or market (i.e., not listed on the
NYSE, AMEX, or Nasdaq National Market). All other offerings are ordinary IPOs. Offer price is the
issue’s offer price. Offer size is the gross proceeds of the offer in millions of dollars, not including the
over-allotment option, and is not inflation-adjusted. HT is a dummy variable equal to one if the issue is
classified as high-tech, zero otherwise. VC is a dummy variable equal to one if the issue is venture capital-
backed, zero otherwise. CM rank is the Carter and Manaster (1990) underwriter prestige ranking for the
lead underwriter (as updated by Loughran and Ritter, 2004). Enforce is a dummy variable equal to one if
the lead underwriter is one of a group of 39 that were the subject of SEC enforcement actions and/or other
penalties during the 1990s and early 2000s, zero otherwise. Initial return is the percentage return from the
offer price to the closing price on the first day of trading. Lockup is the length of the lockup period in
days. Spread is the reported gross dollar spread (i.e., underwriting discount) paid to the underwriter as a
percentage of the offer size, not including nonaccountable expense allowances. 3YrRet is the total
dividend- and split-adjusted buy-and-hold return in percent, calculated from the end of the first trading
day to three years (i.e., 756 trading days) after the first day of trading (or the delisting return date,
whichever is earlier). 3YrMARet is equal to 3YrRet less the contemporaneous compounded return on the
CRSP Nasdaq value-weighted market index. We obtain our data from Thomson Financial’s SDC New
Issues database and the CRSP database. We obtain p-values from a t-test of equality of means between
penny stock and ordinary IPOs, which we calculate by assuming equal or unequal variances, depending
on a test of equality of variances. We assume independence of observations. The sample period is 1990–
1998. We exclude firms not identified on the CRSP database, limited partnerships, real estate investment
trusts, spin-offs, reverse LBOs, mutual-to-stock conversions, American/global depositary receipts, units
and unit trusts, shares of beneficial interests, and firms incorporated outside of the U.S., which results in a
total of 251 penny stock and 2,707 ordinary IPOs.

                                                                                               Penny Stock
                                                                                               Compared to
                                 Penny Stock IPOs                    Ordinary IPOs                Ordinary
 Variable                        N     Mean Std dev                 N     Mean Std dev             p-value
 Offer price, $                251        4.42       0.80       2,707      11.83        4.12          0.0001
 Offer size, $ millions        251         5.7         2.3      2,707       43.6        58.7          0.0001
 HT                            251        0.32       0.47       2,707       0.42        0.49          0.0026
 VC                            251        0.16       0.36       2,707       0.44        0.50          0.0001
 CM rank                       251         2.6         1.1      2,707        7.1         2.1          0.0001
 Enforce                       251        0.26       0.44       2,707       0.04        0.20          0.0001
 Initial return (%)            251        22.4       29.3       2,707       15.4        25.1          0.0003
 Lockup (days)                 207         452        226       2,297        208        102           0.0001
 Spread (%)                    250         9.7         0.6      2,706        7.2         0.8          0.0001
 3YrRet (%)                    251       -21.7      136.1       2,707       44.4      243.1           0.0001
 3YrMARet (%)                  251      -101.8      136.9       2,707      -39.9      242.8           0.0001




                                                                                                            29
              Table II. Initial Returns: Penny Stock IPOs Compared to Ordinary IPOs

This table presents the mean for Initial return (the percentage return from the offer price to the closing
price on the first day of trading) for penny stock IPOs and ordinary IPOs (first row), and for penny stock
IPOs and ordinary IPOs based on whether the lead underwriter is one of a group of 39 that were the
subject of SEC enforcement actions and/or other penalties during the 1990s and early 2000s (second and
third rows). We label these underwriters as “sanctioned,” and all others as “nonsanctioned.” We define
penny stock IPOs as offerings that are not issued by an investment advisor (e.g., not a closed-end fund),
have an offer price at or below $5, and are not listed on a national exchange or market (i.e., not listed on
the NYSE, AMEX, or Nasdaq National Market). All other offerings are ordinary IPOs. We obtain our
data from Thomson Financial’s SDC New Issues database and the CRSP database. We obtain p-values in
the last column from a t-test of equality of means between penny stock and ordinary IPOs. We obtain p-
values in the last row from a t-test of equality of means between IPOs with sanctioned and nonsanctioned
underwriters. We obtain the p-value in the lower right-hand corner from a t-test of equality of means
between penny stock/sanctioned IPOs and ordinary/nonsanctioned IPOs. We calculate the p-values by
assuming equal or unequal variances, depending on tests of equality of variances. We assume
independence of observations. The sample period is 1990–1998. We exclude firms not identified on the
CRSP database, limited partnerships, real estate investment trusts, spin-offs, reverse LBOs, mutual-to-
stock conversions, American/global depositary receipts, units and unit trusts, shares of beneficial
interests, and firms incorporated outside of the U.S., which results in a total of 251 penny stock and 2,707
ordinary IPOs.

                                                                                              Penny Stock
                                                                                              Compared to
                                                                                                 Ordinary
Initial return (%)                       Penny Stock IPOs           Ordinary IPOs                p-values
                                                       (1)                     (2)                     (3)
All IPOs                        (1)                   22.4                    15.4                 0.0003
                                                (N = 251)             (N = 2,707)

Sanctioned                      (2)                    31.6                   13.3                  0.0011
                                                   (N = 65)              (N = 118)

Nonsanctioned                   (3)                   19.2                    15.5                  0.0522
                                                 (N = 186)             (N = 2,589)

Sanctioned Compared to
Nonsanctioned p-value           (4)                 0.0228                  0.2737                  0.0023




                                                                                                         30
                            Table III. Regression Results for Initial Returns

This table provides regression results, with p-values in parentheses, for initial returns as follows:

Initial return = α + β1PS + β2Enforce + β3PS × Enforce + β4VC + β5VC × Enforce + β6Over + β7HT
          + β8MedRank + β9HighRank + β10Secondary + β11LnSize + β12Integer + β13IPOLag
          + β14NasLag + β15Multiple + β16PartU + β17PartD + β18-25Year Dummies + ε
where

Initial return       = percentage return from the offer price to the closing price on the first day of trading
PS                   = dummy variable equal to one if the issue is a penny stock, zero otherwise
Enforce              = dummy variable equal to one if the lead underwriter is one of a group of 39 that
                         were the subject of SEC enforcement actions and/or other penalties during the
                         1990s or early 2000s, zero otherwise
PS × Enforce         = interaction of PS and Enforce
VC                   = dummy variable equal to one if the issue is VC-backed, zero otherwise
VC × Enforce         = interaction of VC and Enforce
Over                 = “overhang,” which we measure as pre-IPO shares retained divided by total shares
                         offered
HT                   = dummy variable equal to one if the issuing firm is high-tech, zero otherwise
MedRank              = dummy variable equal to one if the lead underwriter has a Carter and Manaster
                         (1990) rank greater than or equal to six and less than eight, zero otherwise
HighRank             = dummy variable equal to one if the lead underwriter has a Carter and Manaster
                         (1990) rank greater than or equal to eight, zero otherwise
Secondary            = percentage of shares offered that are secondary shares
LnSize               = natural logarithm of the CPI-adjusted (1982-1984=100) offer size in millions of
                         dollars
Integer              = dummy variable equal to one if the offer price is an integer, zero otherwise
IPOLag               = an underpricing index defined as the average initial return, in percent, for all IPOs
                         in the sample for the month before the issue date
NasLag               = compounded return, in percent, for the CRSP Nasdaq value-weighted market index
                         for the 21 trading days (i.e., one month) prior to the issue date
Multiple             = dummy variable equal to one if the firm has more than one class of common shares,
                         zero otherwise
PartU                = percentage difference between the offer price and the original mid-file price if the
                         adjustment is positive, zero otherwise
PartD                = percentage difference between the offer price and the original mid-file price if the
                         adjustment is negative, zero otherwise, and,
Year Dummies         = eight dummy variables equal to one if the offering is in year 1991, … , 1998, zero
                        otherwise

Regressions (1) through (5) examine all IPOs over the period. Regression (6) uses only ordinary IPOs,
and regression (7) uses only penny stock IPOs. Column (8) reports the p-value for a difference test
between the coefficients from regressions (6) and (7). We define penny stock IPOs as offerings that are
not issued by an investment advisor (e.g., not a closed-end fund), have an offer price at or below $5, and
are not listed on a national exchange or market (i.e., not listed on the NYSE, AMEX, or Nasdaq National
Market). All other offerings are ordinary IPOs. We obtain our data from Thomson Financial’s SDC New
Issues database and the CRSP database. The sample period is 1990–1998. We exclude firms not identified
on the CRSP database, limited partnerships, real estate investment trusts, spin-offs, reverse LBOs,
mutual-to-stock conversions, American/global depositary receipts, units and unit trusts, shares of
beneficial interests, and firms incorporated outside of the U.S., which results in a total of 251 penny stock
and 2,707 ordinary IPOs.

                                                                                                           31
                                             Table III - Continued

   Dependent variable = Initial return

                                                                                                      p-value:
                                                                                         Penny          (6)
                                                                            Ordinary     Stock       Compared
                                         All IPOs                            IPOs        IPOs          to (7)
                   (1)         (2)          (3)         (4)        (5)         (6)         (7)           (8)
    Intercept       15.4        15.3          15.5       14.0        -0.5         -0.7         3.1       0.7008
                (0.0001)    (0.0001)     (0.0001)    (0.0001)   (0.8635)     (0.8180)    (0.7958)
          PS         7.0         6.5           3.7        4.7         6.0          Na          Na           Na
                (0.0001)    (0.0002)     (0.0549)    (0.0161)   (0.0032)
     Enforce                     2.1          -2.2        2.3         2.6         2.1        15.4       0.0032
                            (0.2882)     (0.3538)    (0.3774)   (0.2867)     (0.3885)    (0.0013)
 PS × Enforce                                 14.6       12.8        10.7         Na          Na            Na
                                         (0.0009)    (0.0035)   (0.0069)
          VC                                              3.4        -1.2         -1.2       -2.7       0.7682
                                                     (0.0006)   (0.2354)     (0.2036)    (0.6601)
 VC × Enforce                                           -19.5       -11.0         -7.3      -21.1       0.2022
                                                     (0.0001)   (0.0138)     (0.1594)    (0.0799)
        Over                                                          2.0          1.9        4.9       0.0307
                                                                (0.0001)     (0.0001)    (0.0053)
          HT                                                          3.5          3.9       -0.8       0.1564
                                                                (0.0002)     (0.0001)    (0.8355)
    MedRank                                                          -0.7         -0.7        Na            Na
                                                                (0.6556)     (0.6132)
   HighRank                                                           1.4          1.4         Na           Na
                                                                (0.3585)     (0.3639)
   Secondary                                                         -0.4         -0.5      -22.2       0.2243
                                                                (0.8692)     (0.8392)    (0.3243)
      LnSize                                                         -1.5         -1.4        -4.6      0.5074
                                                                (0.0318)     (0.0349)    (0.4451)
      Integer                                                         2.9          3.2         0.2      0.4204
                                                                (0.0028)     (0.0015)    (0.9619)
     IPOLag                                                           0.3          0.2         0.9      0.0105
                                                                (0.0001)     (0.0001)    (0.0069)
     NasLag                                                           0.5          0.5        -0.0      0.1460
                                                                (0.0001)     (0.0001)    (0.9885)
     Multiple                                                        -1.4         -1.5       25.1       0.0983
                                                                (0.4177)     (0.3709)    (0.2174)
       PartU                                                          0.8          0.8         0.3      0.0104
                                                                (0.0001)     (0.0001)    (0.2605)
       PartD                                                          0.2          0.3        -0.0      0.1405
                                                                (0.0001)     (0.0001)    (0.9541)
Year Dummies                                                    Included     Included    Included     Included

      Adj R2     0.0055      0.0055       0.0089      0.0156     0.2583      0.2785       0.1104
          N       2,958       2,958        2,958       2,958      2,947       2,699          248




                                                                                                       32
            Table IV. Long-run Returns: Penny Stock IPOs Compared to Ordinary IPOs

The table presents long-run returns for penny stock IPOs and ordinary IPOs (rows 1–5), and for penny
stock IPOs and ordinary IPOs based on whether the lead underwriter is one of a group of 39 that were the
subject of SEC enforcement actions and/or other penalties during the 1990s and early 2000s (rows 6–15).
We label these underwriters as “sanctioned” and all others as “nonsanctioned.” We define penny stock
IPOs as offerings that are not issued by an investment advisor (e.g., not a closed-end fund), have an offer
price at or below $5, and are not listed on a national exchange or market (i.e., not listed on the NYSE,
AMEX, or Nasdaq National Market). All other offerings are ordinary IPOs. We obtain our data from
Thomson Financial’s SDC New Issues database and the CRSP database. 3YrMARet (5YrMARet) is the
total dividend- and split-adjusted buy-and-hold return in percent, calculated from the end of the first
trading day to 756 trading days, i.e., three years, (1,260 trading days, i.e., five years) after the first day of
trading (or the delisting return date, whichever is earlier) minus the contemporaneous compounded return
on the CRSP Nasdaq value-weighted market index. 3YrDelist (5YrDelist) is a dummy variable equal to
one if the firm delists from the NYSE, AMEX, or Nasdaq within three years (five years) after the offer
date due to liquidation (i.e., CRSP delist codes 400-490) or poor performance (i.e., CRSP delist codes 500
and 520-591), zero otherwise. FF alpha is the intercept from the Fama and French (1993) time-series
regression of NetReturn (the monthly value-weighted returns from a portfolio of IPO stocks less the risk-
free rate) against Mkt (the excess value-weighted total market return), SMB (the return from a portfolio of
small capitalization less big capitalization stocks), and HML (the return from a portfolio of high minus
low book-to-market stocks). Market returns, risk-free returns, and returns for the SMB and HML
portfolios are from Ken French’s website (http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/) from
1990 to 2003. The portfolio of IPO stocks used to compute NetReturn comprises all IPOs issued during
the previous 60 calendar months (not including the month of issue) that are still CRSP-listed. We exclude
a month if NetReturn for that month is computed with fewer than ten stocks. We obtain the p-values in
rows 1-15 of the fifth column for 3YrMARet, 5YrMARet, 3YrDelist, and 5YrDelist from a t-test of equality
of means between penny stock and ordinary IPOs. The p-value corresponding to FF alpha is the p-value
from our test of whether the intercept is different from zero in the Fama and French (1993) regression in
which the dependent variable, NetReturn, is from a portfolio that is long in ordinary IPOs and short in
penny stock IPOs. We obtain the p-values in rows 16-20 of columns one and three for 3YrMARet,
5YrMARet, 3YrDelist, and 5YrDelist from a t-test of equality of means between IPOs with sanctioned and
nonsanctioned underwriters. The p-value corresponding to FF alpha is the p-value from our test of
whether the intercept is different from zero in the Fama and French (1993) regression in which the
dependent variable, NetReturn, is from a portfolio that is long in IPOs with nonsanctioned underwriters
and short in IPOs with sanctioned underwriters. We obtain the p-values in rows 16-20 of the fifth column
for 3YrMARet, 5YrMARet, 3YrDelist, and 5YrDelist from a t-test of equality of means between penny
stock/sanctioned IPOs and ordinary/nonsanctioned IPOs. The p-value corresponding to FF alpha is the p-
value from our test of whether the intercept is different from zero in the Fama and French (1993)
regression in which the dependent variable, NetReturn, is from a portfolio that is long in
ordinary/nonsanctioned IPOs and short in penny stock/sanctioned IPOs. We calculate the p-values from
the tests of differences in means by assuming equal or unequal variances, depending on a test of equality
of variances. N is the number of IPOs in t-tests of equality of means and is the number of calendar months
for Fama and French (1993) alphas. We assume independence of observations in our two-sample t-tests.
The sample period is 1990–1998. We exclude firms not identified on the CRSP database, limited
partnerships, real estate investment trusts, spin-offs, reverse LBOs, mutual-to-stock conversions,
American/global depositary receipts, units and unit trusts, shares of beneficial interests, and firms
incorporated outside of the U.S., which results in a total of 251 penny stock and 2,707 ordinary IPOs.




                                                                                                              33
                                        Table IV – Continued

                                                                                     Penny Stock
                                                                                    Compared to
                            Penny Stock***      N        Ordinary***      N      Ordinary p-value
                                    (1) ***    (2)            (3)***     (4)                  (5)
      All IPOs
          3YrMARet (1)          -101.8%***    251         -39.9%***    2,707               0.0001
          5YrMARet (2)          -126.4%***    251         -55.2%***    2,707               0.0004
           3YrDelist (3)          31.5%***    251           6.4%***    2,707               0.0001
           5YrDelist (4)          51.4%***    251          14.3%***    2,707               0.0001
           FF alpha (5)          -1.69%***    139         -0.01%***      164               0.0136

      Sanctioned
         3YrMARet (6)           -131.8%***     65         -97.7%***     118                0.0779
         5YrMARet (7)           -162.6%***     65        -124.8%***     118                0.1007
          3YrDelist (8)           26.2%***     65          18.6%***     118                0.2375
          5YrDelist (9)           44.6%***     65          35.6%***     118                0.2330
         FF alpha (10)           -2.73%***     99          0.01%***     116                0.0924

      Nonsanctioned
       3YrMARet (11)             -91.3%***    186         -37.3%***    2,589               0.0001
       5YrMARet (12)            -113.8%***    186         -52.0%***    2,589               0.0168
        3YrDelist (13)            33.3%***    186           5.8%***    2,589               0.0001
        5YrDelist (14)            53.8%***    186          13.4%***    2,589               0.0001
        FF alpha (15)            -1.33%***    137          0.02%***      162               0.0746

      Sanctioned Compared to Nonsanctioned p-value
        3YrMARet (16)          0.0218***                   0.0001***                       0.0001
        5YrMARet (17)          0.0896***                   0.0002***                       0.0001
         3YrDelist (18)        0.2852***                   0.0006***                       0.0005
         5YrDelist (19)        0.2055***                   0.0001***                       0.0001
         FF alpha (20)         0.2211***                   0.7991***                       0.0151
*** ** *
  , , indicate that the mean for 3YrMARet, 5YrMARet, or FF alpha is statistically different from zero at
the 0.01, 0.05, or 0.10 levels, respectively. A comparable test for 3YrDelist and 5YrDelist is not
applicable.




                                                                                                     34
           Table V. Initial and Long-run Returns: Offer Price Compared to Stock Market

This table presents mean initial and long-run returns based on the IPO’s offer price (above $5 compared
to less than or equal to $5) and the exchange or market where the IPO starts trading (NYSE, AMEX,
Nasdaq National Market, or Nasdaq SmallCap Market). Initial return is the percentage return from the
offer price to the closing price on the first day of trading. 3YrMARet (5YrMARet) is the total dividend-
and split-adjusted buy-and-hold return in percent, calculated from the end of the first trading day to 756
trading days, i.e., three years, (1,260 trading days, i.e., five years) after the first day of trading (or the
delisting return date, whichever is earlier) minus the contemporaneous compounded return on the CRSP
Nasdaq value-weighted market index. 3YrDelist (5YrDelist) is a dummy variable equal to one if the firm
delists from the NYSE, AMEX, or Nasdaq within three years (five years) after the offer date due to
liquidation (i.e., CRSP delist codes 400-490) or poor performance (i.e., CRSP delist codes 500 and 520-
591), zero otherwise. We obtain the p-values in rows 7-23 of the third column from a t-test of equality of
means between IPOs with an offer price above $5 compared to IPOs with an offer price less than or equal
to $5. We obtain the p-values in rows 25-29 in columns one and two from a t-test of equality of means
between IPOs that are initially listed on the Nasdaq National Market compared to the Nasdaq SmallCap
Market. We obtain the p-values in rows 25-29 of the third column from a t-test of equality of means
between IPOs with an offer price greater than $5 and listed on the Nasdaq National Market compared to
IPOs with an offer price less than or equal to $5 and listed on the Nasdaq SmallCap Market. We calculate
p-values by assuming equal or unequal variances, depending on a test of equality of variances. We
assume independence of observations. The sample period is 1990–1998. We obtain our data from
Thomson Financial’s SDC New Issues database and the CRSP database. We exclude firms not identified
on the CRSP database, limited partnerships, real estate investment trusts, spin-offs, reverse LBOs,
mutual-to-stock conversions, American/global depositary receipts, units and unit trusts, shares of
beneficial interests, and firms incorporated outside of the U.S., which results in a total of 2,958 IPOs.




                                                                                                           35
                                            Table V – Continued

                                                                                   Offer price > $5
                                                                                 Compared to. ≤ $5
                                      Offer Price > $5        Offer Price ≤ $5              p-value
                                                   (1)                     (2)                  (3)
 NYSE
          Initial return (1)                   11.2%                      NA                   NA
             3YrMARet (2)                     -52.0%                      NA                   NA
             5YrMARet (3)                     -53.1%                      NA                   NA
              3YrDelist (4)                     2.4%                      NA                   NA
              5YrDelist (5)                     9.3%                      NA                   NA
                      N (6)                      246                       0
 AMEX
          Initial return (7)                    5.0%                    8.8%                0.3612
             3YrMARet (8)                    -82.6%                  -162.1%                0.0019
             5YrMARet (9)                    -65.6%                  -218.0%                0.0407
             3YrDelist (10)                   10.0%                    11.1%                0.9209
             5YrDelist (11)                   20.0%                    22.2%                0.8814
                     N (12)                        50                      9
 Nasdaq National
         Initial return (13)                  16.0%                    21.3%                0.1637
            3YrMARet (14)                    -33.6%                   -36.4%                0.9409
            5YrMARet (15)                    -48.4%                  -125.0%                0.0066
             3YrDelist (16)                     5.3%                   12.5%                0.1468
             5YrDelist (17)                   12.5%                    35.4%                0.0020
                     N (18)                     2187                       48
 Nasdaq SmallCap
         Initial return (19)                  16.3%                    22.4%                0.0249
            3YrMARet (20)                    -86.4%                  -101.8%                0.3538
            5YrMARet (21)                   -114.3%                  -126.4%                0.6048
             3YrDelist (22)                   22.2%                    31.5%                0.0372
             5YrDelist (23)                   37.7%                    51.4%                0.0059
                     N (24)                      167                     251
Nasdaq Natl. Compared to Nasdaq SmallCap p-value
         Initial return (25)                  0.8776                   0.8057               0.0010
            3YrMARet (26)                     0.0006                   0.0925               0.0001
            5YrMARet (27)                     0.0002                   0.9635               0.0002
             3YrDelist (28)                   0.0001                   0.0012               0.0001
             5YrDelist (29)                   0.0001                   0.0426               0.0001




                                                                                                 36
                            Table VI. Lockup Times for Penny Stock IPOs

This table shows the distribution of lockup times for penny stock IPOs. We also compare lockup times for
penny stock IPOs that are VC-backed compared to those that are not VC-backed. We define penny stock
IPOs as offerings that are not issued by an investment advisor (e.g., not a closed-end fund), have an offer
price at or below $5, and are not listed on a national exchange or market (i.e., not listed on the NYSE,
AMEX, or Nasdaq National Market). We obtain our data from Thomson Financial’s SDC New Issues
database and the CRSP database. The sample period is 1990–1998. We exclude firms not identified on the
CRSP database, limited partnerships, real estate investment trusts, spin-offs, reverse LBOs, mutual-to-
stock conversions, American/global depositary receipts, units and unit trusts, shares of beneficial
interests, and firms incorporated outside of the U.S., which results in a total of 251 penny stock IPOs.

                         VC-backed                     Non-VC-backed                    Combined
     Days in          Firms   Cumulative              Firms  Cumulative              Firms   Cumulative
     lockup                    percentage                     percentage                      percentage
        <90               0           0.0                 1          0.6                 1           0.5
         90               1           2.6                 9          5.9                10           5.3
     91-179               2           7.9                 2          7.1                 4           7.2
        180               9          31.6                16         16.6                25          19.3
    181-359               2          36.8                 4         18.9                 6          22.2
    360-366               9          60.5                43         44.4                52          47.3
    367-540               8          81.6                37         66.3                45          69.1
    541-549               2          86.8                 8         71.0                10          73.9
    550-719               0          86.8                 0         71.0                 0          73.9
    720-732               4          97.4                44         97.0                48          97.1
       >732               1         100.0                 5        100.0                 6         100.0

                          38          100.0             169           100.0            207           100.0
       Mean              359 days                       473 days                       452 days




                                                                                                        37
                          Figure 1. Distribution of IPOs by Carter and Manaster (1990) Underwriter Rank

The figure provides the percentage distribution of IPOs by Carter and Manaster (1990) rank (as updated
by Loughran and Ritter, 2004), where nine is the highest rank. Penny stock IPOs appear in grey, ordinary
IPOs in black. We define penny stock IPOs as offerings that are not issued by an investment advisor (e.g.,
not a closed-end fund), have an offer price at or below $5, and are not listed on a national exchange or
market (i.e., not listed on the NYSE, AMEX, or Nasdaq National Market). We obtain our data from
Thomson Financial’s SDC New Issues database and the CRSP database. The sample period is 1990–
1998. We exclude firms not identified on the CRSP database, limited partnerships, real estate investment
trusts, spin-offs, reverse LBOs, mutual-to-stock conversions, American/global depositary receipts, units
and unit trusts, shares of beneficial interests, and firms incorporated outside of the U.S., which results in a
total of 251 penny stock IPOs and 2,707 ordinary IPOs.


                    45%



                    40%



                    35%



                    30%
  Percent of IPOs




                    25%



                    20%



                    15%



                    10%



                    5%



                    0%
                             1        2         3        4              5           6             7   8   9
                                                        Carter and Manaster (1990) Rank

                                                     Penny Stock IPOs             Ordinary IPOs




                                                                                                              38
                                         Figure 2. Initial Returns by Year

This figure shows initial returns by year. Initial return is the percentage return from the offer price to the
closing price on the first day of trading. Penny stock IPOs appear in grey, ordinary IPOs in black. We
define penny stock IPOs as offerings that are not issued by an investment advisor (e.g., not a closed-end
fund), have an offer price at or below $5, and are not listed on a national exchange or market (i.e., not
listed on the NYSE, AMEX, or Nasdaq National Market). We obtain our data from Thomson Financial’s
SDC New Issues database and the CRSP database. The sample period is 1990–1998. We exclude firms
not identified on the CRSP database, limited partnerships, real estate investment trusts, spin-offs, reverse
LBOs, mutual-to-stock conversions, American/global depositary receipts, units and unit trusts, shares of
beneficial interests, and firms incorporated outside of the U.S., which results in a total of 251 penny stock
IPOs and 2,707 ordinary IPOs.



                   40%



                   35%



                   30%



                   25%
  Initial Return




                   20%



                   15%



                   10%



                   5%



                   0%
                         1990   1991   1992        1993     1994        1995   1996     1997      1998
                                                            Year


                                         Penny Stock IPOs      Ordinary IPOs




                                                                                                           39

				
DOCUMENT INFO