ssrn-id1433465
DESCRIPTION
LigonTRANSCRIPT
-
Electronic copy available at: http://ssrn.com/abstract=1433465
The Effect of the Trading System on IPO Underpricing: Evidence from the 1997 Order-Handling Rules
James A. Ligon Department of Economics, Finance & Legal Studies
The University of Alabama P.O. Box 870224
Tuscaloosa, AL 35487 Phone: 205-348-6313
Hao-Chen Liu Department of Economics and Finance
The College of Charleston 66 George Street
Charleston, SC 29424 Phone: 843-953-1988
We thank Mike Adams, Carol Carroll, Tom Downs, and Brian Gray for comments on prior drafts. We also thank Jay Ritter and the University of Florida for making some of the data publicly available through the Universitys web site. Remaining errors are our own.
November 20, 2008
-
Electronic copy available at: http://ssrn.com/abstract=1433465
1
The Effect of the Trading System on IPO Underpricing:
Evidence from the 1997 Order-Handling Rules
Abstract
This study uses a natural experiment resulting from the 1997 Securities and Exchange Commission rule mandating a change in the order handling rules (OHR) for all NASDAQ stocks to test whether secondary market structure affects IPO underpricing. We find that the increase in liquidity that the OHR represent led to a decrease in underpricing for cold IPOs suggesting that the expected cost of underwriter price support is a factor in the level of IPO underpricing for cold IPOs, at least on markets where the lead underwriter and principal secondary market maker are usually the same entity.
-
1
The Effect of the Trading System on IPO Underpricing: Evidence from the 1997 Order-Handling Rules
I. Introduction
For initial public offerings listed on the NASDAQ, the lead underwriter usually
becomes the primary market maker in the secondary market (Aggarwal, 2000; Ellis,
Michaely, and OHara, 2000). This differs from the NYSE where the specialist
responsible for making the secondary market in a new issue is normally not affiliated
with the lead underwriter. This raises the question whether the structure of the secondary
market trading system, and in particular whether the fact that in dealer markets like the
NASDAQ the lead underwriter and principal secondary market maker may be the same
entity, could influence aspects of the primary market such as initial public offering (IPO)
underpricing (i.e. the phenomenon that on average the first secondary market closing
price is greater than the offering price). This study uses a natural experiment resulting
from the 1997 Securities and Exchange Commission rule mandating a change in the order
handling rules (OHR) for all NASDAQ issues in order to address this question.
The OHR, which require a limit order to be posted on the trading system if it betters
a dealers quotes, represent a discreet change in the secondary market liquidity of
NASDAQ issues. It is well recognized that the participation of the limit orders on the
NYSE improves market liquidity. Chung, Van Ness, and Van Ness (1999) find that, on
the NYSE, spreads are the narrowest when the amount of limit order trading is the
highest. Corwin, Harris, and Lipson (2004) conclude that limit orders are an important
source of liquidity on the NYSE IPO aftermarket. Before 1997, a NASDAQ dealer was
not required to post a limit order on the trading system even if the limit order prices were
better than the dealers quotes. Thus, limit orders on NASDAQ constituted nothing more
-
2
than proposed terms of trade with the dealer with which they were placed. In the
aftermath of the controversy arising out of the publication of the Christie and Schultz
(1994) study, in 1997 the Securities and Exchange Commission mandated the OHR for
all NASDAQ issues requiring a limit order to be posted on the trading system if it betters
a dealers quotes. Therefore, the OHR allow limit orders to compete directly with dealer
quotes on the NASDAQ and represent a discrete increase in the liquidity of the
NASDAQ market. Overall, studies show that the execution cost differentials across these
two markets narrowed after the market rule changes, consistent with increasing liquidity
(Huang and Stoll, 1996; Bessembinder and Kaufman, 1997; Bessembinder, 1999, 2003;
Barclay, Christie, Harris, Kandel, and Schultz, 1999; Chung and Van Ness, 2001).
Secondary market liquidity and IPO underpricing might be expected to be linked.
Ellul and Pagano (2006) argue that the required returns of initial IPO investors adjust in
response to expected secondary market liquidity, with lower expected secondary market
liquidity leading to higher initial returns. Under their arguments there should be a
negative correlation between expected liquidity and underpricing across the general
cross-section of IPO issues.
Liquidity may also affect initial returns in a narrower way. In IPO aftermarkets
where the underwriter becomes the principal secondary market maker, the underwriter is
concerned with establishing a liquid secondary market for the issue. In order to do so, it is
also well known that if the issue is cold (i.e. demand is low), the underwriter may, and
often does, support the price of the issue in the secondary market by standing ready to
buy shares at the offer price (Lewellen, 2006). This action is called underwriter price
support and is analogous to writing put options with the offer price as the exercise price
-
3
(Hanley, Kumar, and Seguin, 1993; Schultz and Zaman, 1994). Providing these options
to the market is potentially costly to the underwriter if they are exercised. One argument
for the existence of IPO underpricing is that it reduces the costs of underwriter price
support (Schultz and Zaman, 1994). However, if the link between underpricing and
liquidity is primarily associated with underwriter price support, the relationship would be
expected to be strongest whenever the expected cost of price support is highest, which is
when the issue is cold. Under these arguments, changes in liquidity would be expected to
display a negative correlation with underpricing among cold issues.
In this study, we test these arguments by examining IPOs listed on both the
NASDAQ and NYSE both before and after the change in the OHR. Specifically, we
examine whether the discrete change in liquidity implied by the implementation of the
OHR affects the relative underpricing levels of NYSE-eligible NADAQ issues versus
NYSE issues. We consider both the general cross-section and a sub-sample of cold IPOs.
Our results confirm that, for cold IPOs, underpricing on the NYSE was marginally
significantly smaller than on the NASDAQ prior to the change in the OHR, consistent
with the proposition (and prior evidence) that liquidity on the NYSE was higher than on
NASDAQ. After the change in the OHR, we find that underpricing of NYSE-eligible
NASDAQ issues decreased significantly and that there is then no significant difference in
underpricing between the NYSE and NASDAQ. The significance of these results
becomes stronger when the Internet bubble period is excluded. The results thus provide
evidence that underpricing is related to the expected cost of underwriter price support and
that increases in the expected liquidity of the aftermarket reduce the expected cost of
price support and hence reduce IPO underpricing.
-
4
The literature regarding the connections between market structure and initial public
offerings is not extensive. Corwin and Harris (2001) and Anderson and Dyl (2006)
consider the listing decision of firms choosing between the NYSE and NASDAQ. The
former find the presence of industry peers to be a significant factor, while the latter
consider the rules regulating the sale of restricted stock. Neither study is concerned with
underpricing. Affleck-Graves, Hegde, Miller, and Reilly (1993) study the magnitude of
IPO underpricing under different trading systems and find significant differences in the
level of IPO underpricing between NASDAQ/non-NMS and other markets (i.e. the
NYSE, NASDAQ/NMS, and AMEX) but no significant differences in the level of
underpricing among the NYSE, NASDAQ/NMS, and AMEX. They attribute the finding
to the higher listing standards imposed by the NYSE, NASDAQ/NMS, and AMEX. Our
contribution is thus that we are among the first to consider the relationship between
secondary market microstructure and IPO underpricing.
The rest of the paper is organized as follows. Section II provides the testable
hypothesis. Section III describes the sample and data. Section IV reports the results.
Section V provides additional tests to examine the robustness of the results. Section VI
concludes.
II. Testable Hypothesis
Our testable hypothesis is:
H0: The expected liquidity of the secondary market is unrelated to the level of IPO underpricing. H1A: The expected liquidity of the secondary market is negatively correlated with the level of IPO underpricing in the general cross-section of IPO stocks traded on that market.
-
5
H1B: The expected liquidity of the secondary market is negatively correlated with the level of IPO underpricing in the population of cold IPO stocks traded on that market.
The first alternative hypothesis is based upon the arguments of Ellul and Pagano (2006)
that there should be a negative correlation between the level of expected secondary
market liquidity and IPO underpricing. Alternative hypothesis B is based upon the
arguments of Schultz and Zaman (1994), among others, related to the expected cost of
underwriter price support, specifically, since lower secondary market liquidity should
increase the expected cost of price support, there should be a negative correlation
between expected liquidity of cold IPOs and the level of their underpricing. IPOs listed
on the NYSE serve as a control sample.
We conjecture that in the pre-OHR period, all else equal, a NASDAQ IPO should
be underpriced more than a NYSE IPO to compensate investors for the lower expected
liquidity or the underwriter for the higher expected cost of supporting a NASDAQ IPO.
Since the OHR narrow the expected liquidity difference between the NYSE and
NASDAQ trading environments, the differences in underpricing caused by the
differences in expected liquidity should be narrowed or eliminated in the post-OHR
period. If this should be particularly true for cold IPOs, then the expected cost of
underwriter price support is important for IPO underpricing. If it is true for the general
cross-section, then a general relation between underpricing and expected liquidity is
indicated.
III. Sample Selection and Data Description
The data in this study cover U.S. IPOs from 1993 to 2005. Basic information on
IPO firms comes from the Thomson Financials Securities Data Company (SDC) U.S.
-
6
New Issues Database. IPO firms first day closing price and the number of shares
outstanding come from the Center for Research in Security Prices (CRSP) data base.
Firm age, underwriters ranking, and the list of Internet firms come from Jay Ritters IPO
website at the University of Florida. Following previous IPO studies, we exclude IPOs
that are unit offerings, REITs, close-end funds, foreign issues, rights issues, non-firm-
commitment IPOs, and IPOs with offer prices less than $5. Utility and financial issuers
are also excluded from the study. All dollar values in the study are adjusted to the year
2005 dollar values based on the CPI.
Since the goal of this study is to compare the effect of the trading system on IPO
underpricing, only IPOs that list on either the NYSE or NASDAQ are included. IPOs that
occurred in the year of 1997, which is the year of the implementation of the OHR, are
excluded from the sample. We define the pre-OHR period as the years from 1993 to 1996
and the post-OHR period as the years from 1998 to 2005. The Internet bubble period is
defined as the years from 1999 to 2000 and we present results both including and
excluding the bubble years.
Since fundamental differences exist between NYSE and NASDAQ IPOs in
general, to make these two groups of IPOs more comparable, we include only NASDAQ
IPOs that are qualified to be listed on the NYSE, but choose to issue on the NASDAQ
instead. The NYSE listing requirements are based on the NYSE Fact Book (see Table I).
The pre-IPO financial data come from Mergent and EDGAR. The final sample includes
237 IPOs from 1993 to 2005, excluding 1997.
[Insert Table I]
-
7
Table II compares the IPO firms characteristics and offering structures for the
full sample period before and after the implementation of the OHR across the two
markets and within each market. Panel A presents data on means and Panel B presents
data on medians. Since the Internet bubble period represents an irregular trading period,
we also perform the same statistical comparisons with the Internet bubble period
excluded from the sample (Panels C and D).
[Insert Table II]
A. Across Market Analysis
In this sub-section, we compare the IPO firms differences across markets. We
look at the cross-market differences for the entire sample period and the differences in the
pre-OHR and the post-OHR periods. For the entire sample period, the mean initial return
of the NASDAQ IPOs is significantly higher than that of the NYSE IPOs, while the
means and medians of the offer price, IPO proceeds, and market value of the NASDAQ
IPOs are lower. The mean age of the NASDAQ IPOs is also significantly lower than that
of the NYSE IPOs. With respect to the offering structures, the NYSE and NASDAQ
IPOs do not show much difference. The proportions of hot and cold IPOs are not
significantly different across these two markets. The mean and median retention ratios
are similar across the markets. In the pre-OHR period, the mean and median of the initial
return are still significantly different between these two markets at the 5% level but the
significance disappears in the post-OHR period. Also, the mean firm age and proceeds
are significantly different between these two markets in the post-OHR period. These
results could be driven by the Internet bubble period during the post-OHR period.
-
8
Panels C and D present the same comparisons excluding the Internet bubble
period and the results show that it is indeed the Internet bubble period that contributes
some of the significant differences across the markets in the post-OHR period. When
excluding the Internet bubble period, we do not find any statistically significant
differences in the means of proceeds and firm age across markets in the post-OHR period.
B. Within Market Analysis
Table II further reports the changes of the firms characteristics and the offering
structure for the periods before and after the implementation of the OHR for each market.
When we compare the changes within each market, we find that the characteristics of the
IPOs on both markets do not vary much across time. For the NYSE IPOs, the only
significant changes are offer price, proceeds, and market value. For the NASDAQ IPOs,
in the post-OHR period, the mean and median of offer price are lower while the medians
of proceeds and market value are higher.
IV. Analysis and Results
The objective of this study is to test whether the trading system has any effect on
IPO underpricing. Underpricing is defined as:
ii
ii
i
ii
firmfor priceoffer IPO theis PriceOffer firmfor day ingfirst trad on the pricelast theis Price Closing
firmsfor index an is i
(1) PriceOffer
PriceOffer Price Closing
where
ngUnderprici i
=
We pool both the NYSE and NASDAQ IPOs and use the market- and rule-related
dummy variables to examine how the OHR affect the differential in the level of
-
9
underpricing within each market and across markets. From the previous section, we see
that the IPOs issued on the different markets show fundamental differences. Therefore, it
is important to control for the differences in the IPO firms characteristics and the
offering structures in order to see the effect of the market trading system on the level of
underpricing. To control for the systematic differences for the IPOs issued on these two
markets, following the previous studies on the IPO underpricing, we include several
control variables common to IPO underpricing studies (see Wang and Ligon (2008) and
Jones and Ligon (2008) for example). The regression model is as follows (see Appendix
A for the description of each variable):
(2) eePriceUpdataIntegeraRetentionaLockupDaysa
Ranking UWaVCaTech-HiaInterneta1)(Agea alue)ln(MarketVaRuleaNasdaq*RuleaNasdaqaangUnderprici
i141311
109865
4210i
+++++++++++
++++=
12
3
In equation (2), we test the market and rule effects on the level of underpricing by
including a market dummy variable, Nasdaq, and a OHR dummy variable, Rule. Nasdaq
equals 1 if the IPO is issued on the NASDAQ and 0 otherwise and Rule equals 1 if the
IPO is issued after year 1997 and 0 otherwise. To account for the effect of the rule
changes on the NASDAQ IPOs in the post-OHR period, we include an interaction term,
Rule*Nasdaq. To control for the firms characteristics, we include market value, age,
Internet and Hi-Tech dummies (including industry dummies produces the same results),
and a venture capital backed dummy (VC). To control for the offering structure, we
include the underwriters quality ranking (UW Ranking), the lockup period (Lockup
Days), the percent of shares retained by the pre-IPO shareholders (Retention), a dummy
variable for integer price offerings (Integer), and the percentage price adjustment
(PriceUpdate).
-
10
We repeat the analysis for a sub-sample of cold IPOs. A cold IPO is defined as
one with the offer price lower than the bottom of the filing range. We also test the effects
on underpricing with and without the Internet bubble period. For all models, we check for
multicollinearity and do not find any substantial issues. We also correct for
heteroscedasticity by using robust estimators in all models. We present the regression
results in Tables III and V, and the effects of market system and rule change on the
underpricing are summarized in Tables IV and VI for the full sample and the cold IPO
sub-sample, respectively.
A. Full Sample
The regression results for the full sample IPOs are reported in Table III and the
effect of the OHR on underpricing is summarized in Table IV. In Table III, model (1) is
for the entire sample period and model (2) is the one without the Internet bubble period.
To see the effect of the rule change on the level of IPO underpricing, we focus on three
dummy variables, Nasdaq, Rule*Nasdaq, and Rule, which can be used to compare the
effect of the OHR on the level of underpricing across markets and within each market in
the pre-OHR and the post-OHR periods. For the within market effect of the OHR, we test
if the coefficient on Rule is significant for the NYSE IPOs and the coefficients on
Rule*Nasdaq and Rule are jointly significant for the NASDAQ IPOs. For the cross-
market effect of the OHR, in the pre-OHR period, we test if the coefficient on Nasdaq is
significant. In the post-OHR period, we test if the coefficients on Nasdaq and
Rule*Nasdaq are jointly significant. Table IV reports the summary of the test results on
the market and the rule change effects on the level of underpricing.
[Insert Tables III and IV]
-
11
In Table III, the coefficients of the interaction term, Rule*Nasdaq, have the
expected negative sign in both models but are not statistically significant. In Table IV,
both panels A and B (with and without the Internet bubble period) show no significant
differences in the level of underpricing for the periods before and after the OHR within
each market and across markets. Thus, we cannot reject our null hypothesis in favor of
alternative hypothesis A for the full sample.
B. Cold IPOs
When we examine the effect of the OHR on the cold IPOs, we see a different
story. In Table V, the coefficients of the interaction term, Rule*Nasdaq, also have the
expected negative sign in both models but, in contrast to the full sample, both coefficients
are statistically significant. For the within market effect, Panel A of Table VI shows that
the implementation of the OHR does not have a significant effect on the underpricing
level of the NYSE IPOs when we compare the underpricing level before and after the
OHR. This finding is not surprising for the NYSE IPOs since the OHR affect only the
NASDAQ issues. NASDAQ IPOs, on the other hand, are less underpriced in the post-
OHR period and the difference is significant at a level close to 10%. The result is more
robust when we exclude the Internet bubble period (Panel B). In Panel B of Table VI, the
underpricing level of the NASDAQ IPOs is lower in the post-OHR period compared to
the pre-OHR period and the difference is significant at the 5% level. When comparing the
cross market effect, in Panel A, before the OHR, the underpricing level of the NASDAQ
IPOs is significantly higher at a level close to 10% but the difference in the underpricing
level disappears between these two markets in the post-OHR period. When we exclude
the Internet bubble period, in Panel B, the result is more robust. In the pre-OHR period,
-
12
the NASDAQ IPOs on average are significantly more underpriced than the NYSE IPOs
at the 5% level. In the post-OHR period, the difference across these two markets
disappears. Therefore, the results provide evidence that the OHR help eliminate the
differences in the level of underpricing between these two markets.
[Insert Tables V and VI]
The results are consistent with our alternative hypothesis B. For the full sample,
there is no real evidence that underpricing on the NASDAQ declined because of the rule.
On the other hand, the results for the cold IPOs show significant differences in the IPO
underpricing caused by the implementation of the OHR in both cases of within the
NASDAQ market and across markets. For the cold IPOs, since the demand is low in the
pre-IPO market, the expected cost of price support would be high in the aftermarket.
Therefore, in the pre-OHR period, since the liquidity was greater on the NYSE than on
the NASDAQ, all else equal, the underwriter would assess the higher expected cost of
price support for a NASDAQ IPO versus an otherwise equivalent NYSE IPO and
therefore underprice the NASDAQ IPO more than the otherwise equivalent NYSE IPO.
Our results confirm that underpricing on the NYSE was significantly smaller than on the
NASDAQ prior to the OHR, consistent with this proposition. After the change in the
OHR rules, limit orders compete freely with dealer bid and ask quotes, suggesting that
the OHR improved liquidity on the NASDAQ. Consistent with this proposition, our
empirical results confirm that after the implementation of the OHR, there is no significant
difference in underpricing between the NYSE issues and NYSE-eligible NASDAQ issues.
The fact that nothing seems to have changed in the general cross-section supports the
argument that the expected cost of price support is driving the results. The results thus
-
13
provide evidence that underpricing is related to the expected cost of underwriter price
support and that actions that increase the expected liquidity of the aftermarket reduce the
expected cost of price support and hence reduce IPO underpricing, at least among cold
issues.
V. Robustness Tests
In this section, we provide additional tests to examine whether there are factors
other than the OHR that drives the results presented in the previous section. Specifically,
we examine if the increase in the use of odd-eighth quotes in response to the Christie and
Schultz (1994) study, which occurred prior to the implementation of the OHR, affects the
results. Christie and Schultz (1994) (CS) document larger bid-ask spreads for NASDAQ
stocks that avoided odd-eighth quotes and conclude that the avoidance of odd-eighth by
NASDAQs market makers is a result of collusion among NASDAQ market makers.
Christie, Harris, and Schultz (1994) document that odd-eighth avoidance was
significantly curtailed immediately upon publication of media reports regarding the
results of CS. The avoidance of odd-eighth quotes contributed partly to the higher
transaction costs on NASDAQ. Barclay, Christie, Harris, Kandel, and Schultz (1999)
compare the size of inside spreads of NASDAQ stocks over 3 time intervals, (1) from
January, 1994 to May, 1994, the period before the CS study, (2) from November 1996 to
January 1997, the period after their study and before the implementation of OHR, and (3)
February 1997, the month immediately after the implementation of OHR. Their results
show that between periods (1) and (2), the average size of spreads dropped significantly
for the stocks that avoided odd-eighth quotes but was unchanged for the stocks that used
both odd- and even-eighth quotes. Therefore, they conclude that the increase in the use of
-
14
odd-eighth quotes has a significant effect on the size of spreads. With respect to the effect
of the OHR on spreads, their study also shows that the OHR have a significant effect on
lowering the size of spreads for NASDAQ stocks. The average size of spreads for both
stocks using and avoiding odd-eighth quotes prior to the CS study dropped significantly
(at the 1% level) further after the implementation of OHR. On the other hand,
Bessembinder and Kaufman (1997), who study a larger sample of NASDAQ do not find
lower quoted and effective spreads in the July-December 1994 period compared to the
January-May 1994 period. Small and medium sized firms show significantly increased
spreads, while large size firms showed insignificantly decreased spreads. They also find
that after May 1994, the frequency of even-eighth rounding declines significantly (except
for the smallest firms), implying that NASDAQ market makers did increase the use odd-
eighth quotes. Taken together, these results suggest that even-eighth rounding might not
be the source of larger NASDAQ spreads.
To test if the avoidance of odd-eighths has any effect on our study, we perform
two analyses. First, we examine the quoted, percentage quoted, realized, and percentage
realized spreads of NASDAQ IPOs using the closing bids, asks, and prices of the IPOs
under different time intervals. Following Barclay et al. (1999), we separate our sample
into 3 sub-sample periods, (1) from January 1993 to May 1994, (2) from May 1994 to
December 1996, and (3) from January 1998 to April 2001 where April 2001 is the month
of decimalization on NASDAQ. The spread measures are defined as follows:
-
15
( )100
Midpoint QuotedSpread Realized Spread Realized Percentage )4(
Midpoint Quoted - Price ClosingDaily 2 Spread Realized )3(2
Bid Quoted Ask Quoted Midpoint Quoted where
100Midpoint QuotedSpread Quoted Spread Quoted Percentage (2)
BidQuoted-Ask Quoted Spread Quoted )1(
=
=
+=
=
=
Table VII shows these spread measures for NASDAQ stocks in 3 sub-periods. We
examine the entire population of NASDAQ IPOs, NASDAQ IPOs with mixed-eighth
quotes, and NASDAQ IPOs with even-eighth quotes IPOs. An IPO is defined as a mix-
eighth quote IPO if either the closing bid or ask on the date of the IPO was an odd-eighth
and an even-eighth quote IPO if both the closing bid and ask on the date of the IPO were
even-eighth. Consistent with Bessembinder and Kaufman (1997), we find no evidence
that the average sizes of spreads of IPOs issued between May 1994 and before the
implementation of the OHR are lower than the sizes of spreads of IPOs issued prior to
May 1994. However, we do find that OHR lowered the sizes of spreads of NASDAQ
IPOs significantly. Therefore, we do not find evidence that the liquidity changed
significantly during the pre-OHR period.
[Insert Table VII]
Second, to separate the effect of OHR from that of the increase in odd-eighth
quotes prior to OHR, we re-run the same tests in the previous section using 2 sub-periods,
(1) January 1993 to December 1996, and (2) May 1994 to December 2005. To capture
the effect of changes in odd-eighth quotes, we add 2 dummy variables, Odd and
Odd*NASDAQ, in the first sub-period. Odd equals 1 if an IPO was issued after May 1994,
-
16
and 0 otherwise. Odd*NASDAQ is an interaction term which accounts for the effect of
the changes in the increase of odd-eighth quotes on NASDAQ IPOs between May, 1994
and December 1996. The sample includes both the NYSE and NYSE-qualified NASDAQ
IPOs. The Internet bubble period is excluded.
Tables VIII and X present the regression results for the full sample and cold
issues, respectively. The results are similar to the ones presented in the previous section.
We do not find evidence that market structure affects the difference on IPO underpricing
between the NYSE and NASDAQ IPOs for the full sample. For the periods before OHR,
we find that, for the full sample, the NASDAQ IPOs issued between 1994 and 1996 are
more significantly underpriced than the NASDAQ IPOs issued before 1994. When
comparing the levels of underpricing across NYSE and NASDAQ, we do not find any
evidence that the increase in the use of odd-eighth quotes or the implementation of OHR
has any significant effect on IPO underpricing for the full sample. (see Tables IX)
[Insert Tables VIII and IX here]
With respect to the within market effect on the cold issues prior to 1994, since
there was only one NASDAQ cold issue during the pre-1994 period, we only include
Odd and Odd*NASDAQ and omit the NASDAQ dummy in order to capture the effect
resulting from the change in odd-eighth quotes. Table XI, Panel A shows that the change
in the use of the odd-eighth quotes on NASDAQ does not have any significant effect on
IPO underpricing before the implementation of the OHR. However, Panel B of Table XI
shows that when we examine the effect of the OHR on cold IPO underpricing, NASDAQ
IPOs were significantly less underpriced after the implementation of OHR. When we
compare the difference on the levels of underpricing across markets, we also find that
-
17
before the OHR, NASDAQ cold IPOs were significantly more underpriced than NYSE
IPOs, but after the implementation of the OHR, the difference caused by the difference in
market structure disappeared. Therefore, we do not find evidence that the difference in
the level of underpricing between the NYSE and NASDAQ cold IPOs was narrowed due
to the increase in the odd-eighth quotes prior to OHR.
[Insert Tables X and XI here]
VI. Conclusion
The purpose of this paper is to study the effect of the trading system on the level
of the IPO underpricing. We compare the effect of the 1997 change in the NASDAQ
order-handling rules on underpricing across the NYSE and NASDAQ IPOs. The change
in the order handling rules improved liquidity on the NASDAQ. Our results support the
argument that the expected cost of price support in the aftermarket is assessed ex ante by
the lead underwriter when determining the IPO offer price. Any event that exogenously
improves the liquidity of the secondary market would lower the expected cost of price
support, and hence, underpricing, at least among cold issues. Therefore, underpricing and
the expected cost of price support (aftermarket liquidity) are positively (negatively)
correlated. Our results show that before the OHR, the NYSE-eligible NASDAQ IPOs
were significantly more underpriced than the NYSE IPOs while after the OHR, there is
no significant difference. Therefore, the results provide evidence that the trading system
plays a role in IPO underpricing.
-
18
References Affleck-Graves, J., S.P. Hegde, R.E. Miller, and F.K. Reilly, 1993, The Effect of the Trading System on the Underpricing of Initial Public Offerings, Financial Management 22, 99-108. Aggarwal, R, 2000, Stabilization Activities by Underwriters after Initial Public
Offerings, Journal of Finance 55, 1075-1102. Anderson, A.A. and E.A. Dyl, 2008, IPO Listings: Where and Why?, Financial
Management 37, 23-43. Barclay, M.J., W.G. Christie, J.H. Harris, E. Kandel, and P.H. Schultz, 1999, Effects of
Market Reform on the Trading Costs and Depth of NASDAQ Stocks, Journal of Finance 54, 1-34.
Bessembinder, H., 1999, Trade Execution Costs on NASDAQ and the NYSE: A Post- Reform Comparison, Journal of Financial and Quantitative Analysis 34, 387- 407. Bessembinder, H., 2003, Trade Execution Costs and Market Quality after
Decimalization, Journal of Financial and Quantitative Analysis 38, 747-777. Bessembinder, H. and H.M. Kaufman, 1997, Comparison of Trade Execution Costs for
NYSE and NASDAQ-Listed Stocks, Journal of Financial and Quantitative Analysis 32, 287-310.
Christie, W.G. and P.H. Schultz, 1994, Why Do NASDAQ Market Makers Avoid Odd-
Eighth Quotes?, Journal of Finance 49, 1813-1840. Christie, W.G., J.H. Harris, and P.H. Schultz, 1994, Why Did NASDAQ Market Makers
Stop Avoiding Odd-Eighth Quotes?, Journal of Finance 49, 1841-1860. Chung, K.H. and R.A. Van Ness, 2001, Order Handling Rules, Tick Size, and the
Intraday Pattern of Bid-Ask Spreads for NASDAQ Stocks, Journal of Financial Markets 4, 143-161.
Chung, K.H., B. Van Ness, and R. Van Ness, 2002, Spreads, Depths, and Quote
Clustering on the NYSE and NASDAQ: Evidence after 1997 Securities and Exchange Commission Rule Changes, Financial Review 37, 481-505.
Corwin, S.A. and J.H. Harris, 2001, The Initial Listing Decisions of Firms that Go
Public, Financial Management 30, 35-55. Corwin, S.A., J.H. Harris, and M.L. Lipson, 2004, The Development of Secondary Market Liquidity for NYSE-Listed IPOs, Journal of Finance 59, 2339-2373.
-
19
Ellis, K., R. Michaely, and M. OHara, 2000, When the Underwriter Is the Market Maker: An Examination of Trading in the IPO Aftermarket, Journal of Finance 55, 1039-1074. Ellul, A. and M. Pagano, 2006, IPO Underpricing and After-market Liquidity, Review
of Financial Studies 19, 381-421. Hanley, K.W., A.A. Kumar, and P.J. Seguin, 1993, Price Stabilization in the Market for
New Issues, Journal of Financial Economics 34, 177-197. Huang, R.D. and H.R. Stoll, 1996, Dealer versus Auction Markets: A Paired Comparison of Execution Costs on NASDAQ and the NYSE, Journal of Financial Economics 41, 313-357. Jones, T.L. and J.A. Ligon, 2008, The Day of the Week Effect in IPO Initial Returns,
Quarterly Review of Economics and Finance, forthcoming. Lewellen, K., 2006, Risk, Reputation, and IPO Price Support, Journal of Finance 61,
613-653. New York Stock Exchange Fact Book, 1993-2000 Schultz, P.H. and M.A. Zaman, 1994, Aftermarket Support and Underpricing of Initial Public Offerings, Journal of Financial Economics 35, 199-219. Wang, Q. and J.A. Ligon, 2008, The Underpricing of Insurance IPOs, Financial
Management, forthcoming.
-
20
Table I The NYSE listing requirements Table I gives the NYSE IPO listing requirements. The criteria are based on the NYSE Fact Book. (1) Offer proceeds are at least $40 million. (2) Total assets prior to the issue are at least $40 million. (3) Number of shares offered shares is at least 1.1 million shares. (4) Earnings requirements: (a) Pre-tax income in current year is greater than $2.5 million and pre-tax
incomes in the previous two years are at least $2 million each year; or
(b) Pre-tax income in current year is greater than $4.5 million and there are
positive pre-tax incomes in the previous two years prior to the IPO year.
-
21
Table II Comparison of the NYSE-listed IPO firms vs. NYSE-qualified NASDAQ-listed IPO firms Table II reports the means (Panel A) and the medians (Panel B) of IPO firms characteristics and offering structures for IPOs that are listed on the NYSE and IPOs that are qualified to be listed on the NYSE but listed on NASDAQ from 1993 to 2005 excluding 1997. Panels C and D present the same information excluding IPOs from the IPO bubble period (1999-2000). The Overall column is for the entire sample period. The Pre-OHR (Post-OHR) column is for the period before (after) the implementation of the order handling rules (OHR). # of IPOs is the total number of IPOs in the sample, which represents the maximum number of observations for each one of the variables. Initial return is the IPO first-day price change in percentage terms. Offer price is the IPO offering price from SDC. Proceeds are the offering proceeds from SDC, which is in millions of dollars. Cold issue is a dummy variable that equals one if the offering price is lower than the bottom of the filing range. Retention is the proportion of total shares retained by the pre-IPO shareholders. Market value is the total market capitalization of the IPO firm at issuance in millions of dollars. Ranking is the underwriter quality ranking, which ranges from 9.99 (best) to 1 (worst). Age is the firms age at the IPO. VC backed is a dummy variable that equals one if the IPO is backed by venture capital. Both means and medians are compared for the entire period, pre-OHR period, and post-OHR period within each market and across markets. The significance level is indicated by * (at 5% level) and ** (at 1% level) for the within market comparisons and a (at 5% level) and b (at 1% level) for the cross markets comparisons. The significant level of the within-market difference is indicated under the Post-OHR columns and that of the across-market difference is indicated under NASDAQ columns. Full Sample Panel A: Mean NYSE NASDAQ Overall Pre-OHR Post-OHR Overall Pre-OHR Post-OHR # of IPOs 174 82 92 63 27 36 Initial return (%) 9.16 8.47 9.78 15.57a 17.54a 14.08 Offer price 19.61 20.76 18.59** 17.39b 19.75 15.62**b Proceeds (mil) 235.51 134.03 325.97** 130.22a 91.93a 158.93a Cold issue 0.31 0.28 0.34 0.24 0.15 0.31 Retention 0.64 0.64 0.64 0.67 0.67 0.66 Market value (mil) 1003.3 618.06 1346.67** 441.40b 403.86 469.55b Ranking 8.76 8.67 8.83 8.10b 8.10b 8.10b Age 32.15 30.7 33.45 24.32a 26.52 22.67a VC backed 0.21 0.26 0.16 0.32 0.33 0.31
Panel B: Median NYSE NASDAQ Overall Pre-OHR Post-OHR Overall Pre-OHR Post-OHR # of IPOs 174 82 92 63 27 36 Initial return (%) 5.09 4.36 5.77 12.5 12.50a 11.79 Offer price 19.13 20.09 18.00** 17.00b 18.6 16.00**b Proceeds (mil) 155.4 101.39 188.85** 80.78b 72.90b 86.26*b Cold issue 0 0 0 0 0 0 Retention 0.69 0.67 0.7 0.71 0.67 0.71 Market value (mil) 587.53 485.84 663.41** 298.51b 236.96b 359.32*b Ranking 9.1 9.1 9.10* 8.10b 8.10b 8.10b Age 19 17 19 20 17 20.5 VC backed 0 0 0 0 0 0
-
22
Excluding the IPO Bubble Years (i.e. no 1999-2000 offerings) Panel C: Mean NYSE NASDAQ Overall Pre-OHR Post-OHR Overall Pre-OHR Post-OHR # of IPOs 165 82 83 55 27 28 Initial return (%) 9.06 8.47 9.65 14.11a 17.54a 10.79 Offer price 19.67 20.76 18.60** 17.65b 19.75 15.63**b Proceeds (mil) 234.59 134.03 333.94** 136.63a 91.93a 179.73 Cold issue 0.3 0.28 0.31 0.24 0.15 0.32 Retention 0.64 0.64 0.64 0.66 0.67 0.65 Market value (mil) 939.66 618.06 1257.39** 436.37b 403.86 467.72b Ranking 8.75 8.67 8.82 8.10b 8.10b 8.10b Age 31.75 30.7 32.78 25.24 26.52 24 VC backed 0.21 0.26 0.17 0.31 0.33 0.29
Panel D: Median NYSE NASDAQ Overall Pre-OHR Post-OHR Overall Pre-OHR Post-OHR # of IPOs 165 82 83 55 27 28 Initial return (%) 5.21 4.36 5.77 11.76 12.50a 10.04 Offer price 19.2 20.09 18.00** 17.36b 18.6 16.00**a Proceeds (mil) 154 101.39 188.70** 75.37b 72.90b 85.20b Cold issue 0 0 0 0 0 0 Retention 0.69 0.67 0.7 0.67 0.67 0.69 Market value (mil) 584.44 485.84 661.10** 275.47b 236.96b 323.21*b Ranking 9.1 9.1 9.10* 8.10b 8.10b 8.60b Age 19 17 19 21 17 21 VC backed 0 0 0 0 0 0
-
23
Table III Effects of Listing Market and Trading Rules on IPO Underpricing for the NYSE-listed IPOs and the NYSE-qualified NASDAQ-listed IPOs (with bubble period and without bubble period): Full Sample Table III reports the OLS regression estimates of IPO underpricing for the NYSE-listed IPOs and the NYSE-qualified NASDAQ IPOs for the period of 1993 to 2005 excluding 1997. Model (1) includes the bubble period and model (2) excludes the bubble period (i.e. 1999-2000). The dependent variable for both models is the IPO initial return. To correct for heteroscedasticity, the heteroscedasticity consistent estimation method is used in the model. The description of the independent variables is in Appendix A. (1) (2) With Bubble Without Bubble NASDAQ 3.301 4.052 (0.92) (1.22) RULE*NASDAQ -3.988 -6.134 (0.87) (1.41) RULE 2.666 1.803 (1.58) (1.11) LN(MARKET VALUE) -1.443 -0.863 (1.37) (0.79) LN(AGE+1) -0.765 -0.613 (1.36) (1.09) INTERNET 27.067 4.427 (1.85)+ (0.89) HI-TECH -0.303 -0.696 (0.16) (0.37) VC 2.304 1.780 (1.10) (0.85) UW RANKING 0.010 -0.404 (0.01) (0.40) LOCKUP DAYS -0.031 -0.033 (2.09)* (2.11)* RETENTION 12.008 12.218 (3.72)** (3.74)** INTEGER PRICE -0.415 -0.666 (0.26) (0.42) PRICE UPDATE 40.183 40.610 (5.52)** (5.39)** CONSTANT 17.781 18.322 (1.67)+ (1.68)+ Observations 237 220 R-squared 0.39 0.39 Robust t statistics in parentheses + significant at 10%; * significant at 5%; ** significant at 1%
-
24
Table IV Effects of Trading Systems on the Differences in IPO Underpricing within and across Markets (the NYSE-listed IPOs vs. the NYSE-qualified NASDAQ-listed IPOs): Full Sample Table IV reports the test results of the effects of market listing and OHR within each market and across markets for IPOs from 1993 to 2005 excluding 1997. The top panel includes the bubble period and the bottom panel excludes the bubble period (i.e. 1999-2000). The Within Market difference in underpricing is based on post-OHR minus pre-OHR. The Across Market difference in underpricing is based on NASDAQ minus NYSE. Sign is the sign of the joint coefficients from the variables in the null hypothesis. The tests are based on the OLS regression estimates in Table III.
Panel A: With Internet Bubble Period (1) Within Market Null Hypothesis Sign p-value NYSE: (Post-OHR) (Pre-OHR) Rule=0 + 0.1150 NASDAQ: (Post-OHR) (Pre-OHR) Rule*NASDAQ + Rule=0 - 0.7592 (2) Across Market Null Hypothesis Sign p-value Pre-OHR: NASDAQ NYSE NASDAQ=0 + 0.3579 Post-OHR: NASDAQ - NYSE NASDAQ + Rule*NASDAQ=0 - 0.7999
Panel B: Without Internet Bubble Period (1) Within Market Null Hypothesis Sign p-value NYSE: (Post-OHR) (Pre-OHR) Rule=0 + 0.2664 NASDAQ: (Post-OHR) (Pre-OHR) Rule*NASDAQ + Rule=0 - 0.2777
(2) Across Market Null Hypothesis Sign p-value Pre-OHR: NASDAQ NYSE NASDAQ=0 + 0.2234 Post-OHR: NASDAQ - NYSE NASDAQ + Rule*NASDAQ=0 - 0.4193
-
25
Table V Effects of Listing Market and Trading Rules on IPO Underpricing for the NYSE-listed IPOs and the NYSE-qualified NASDAQ-listed IPOs (with bubble period and without bubble period): Cold IPOs Table V reports the OLS regression estimates of Cold IPO underpricing for the NYSE-listed IPOs and the NYSE-qualified NASDAQ IPOs. A Cold IPO is defined as one with the offer price below the bottom of the filing range. Model (1) includes the bubble period and model (2) excludes the bubble period (i.e. 1999-2000). The dependent variable for both models is the IPO initial return. To correct for heteroscedasticity, the heteroscedasticity consistent estimation method is used in the model. The description of the independent variables is in Appendix A. (1) (2) With Bubble Without Bubble NASDAQ 8.052 8.877 (1.64) (2.00)+ RULE*NASDAQ -10.346 -9.620 (1.86)+ (1.95)+ RULE 1.733 0.200 (1.02) (0.14) LN(MARKET VALUE) -0.019 0.308 (0.02) (0.38) LN(AGE+1) -1.037 -0.435 (1.76)+ (0.83) INTERNET 8.488 12.286 (2.39)* (5.98)** HI-TECH 5.002 1.696 (1.56) (0.90) VC -2.726 -2.782 (1.05) (1.08) UW RANKING -0.018 -0.142 (0.02) (0.18) LOCKUP DAYS -0.013 -0.025 (0.59) (1.40) RETENTION 2.966 3.593 (0.79) (0.97) INTEGER PRICE 1.518 1.139 (0.91) (0.85) PRICE UPDATE 11.945 11.575 (1.35) (1.79)+ CONSTANT 5.586 5.532 (0.46) (0.56) Observations 69 62 R-squared 0.32 0.39 Robust t statistics in parentheses + significant at 10%; * significant at 5%; ** significant at 1%
-
26
Table VI - Effects of Trading Systems on the Differences in IPO Underpricing within and across Markets (the NYSE-listed IPOs vs. the NYSE-qualified NASDAQ-listed IPOs): Cold IPOs Table VI reports the test results of the effects of market listing and OHR within each market and across markets for Cold IPOs. A Cold IPO is defined as one with the offer price below the bottom of the filing range. The top panel includes the bubble period and the bottom panel excludes the bubble period (i.e. 1999-2000). The Within Market difference in underpricing is based on post-OHR minus pre-OHR. The Across Market difference in underpricing is based on NASDAQ minus NYSE. Sign is the sign of the joint coefficients from the variables in the null hypothesis. The tests are based on the OLS regression estimates in Table V.
Panel A: With Internet Bubble Period (1) Within Market Null Hypothesis Sign p-value NYSE: (Post-OHR) (Pre-OHR) Rule=0 + 0.3144 NASDAQ: (Post-OHR) (Pre-OHR) Rule*NASDAQ + Rule=0 - 0.1078 (2) Across Market Null Hypothesis Sign p-value Pre-OHR: NASDAQ NYSE NASDAQ=0 + 0.1068 Post-OHR: NASDAQ - NYSE NASDAQ + Rule*NASDAQ=0 - 0.4224
Panel B: Without Internet Bubble Period (1) Within Market Null Hypothesis Sign p-value NYSE: (Post-OHR) (Pre-OHR) Rule=0 + 0.8912 NASDAQ: (Post-OHR) (Pre-OHR) Rule*NASDAQ + Rule=0 - 0.0558
(2) Across Market Null Hypothesis Sign p-value Pre-OHR: NASDAQ NYSE NASDAQ=0 + 0.0513 Post-OHR: NASDAQ - NYSE NASDAQ + Rule*NASDAQ=0 - 0.7300
-
27
Table VII Quoted, percentage quoted, realized, percentage realized spreads for IPOs issued from 1993 to 2001 (excluding 1997) Table VII reports four liquidity measures for NASDAQ IPOs issued between 1993 and 2001 excluding 1997. Panel A reports quoted spread, Panel B reports percentage quoted spread, Panel C reports realized spread, and Panel C reports percentage realized spread. The first time period correspond to the period before the study of Christie and Schultz (1994). The second period corresponds to the period between the Christie and Schultz (1994) study and the implementation of OHR. The third period corresponds to the post-OHR periods. Mixed quotes are the IPOs with either a closing bid or ask at an odd-eighth on the offer date. Even-eighth quotes are the IPOs with both closing bid and ask prices at even-eighths on the offer date. 1/1993 to 5/1994 5/1994 to 12/1996 1/1998 to 4/2001 Panel A: Quoted Spread All 0.3529 0.3666+ 0.0932** Mixed quotes 0.2306 0.2626** 0.1047** Eveneighth quotes 0.3841 0.4181** 0.0814** Panel B: Percentage Quoted Spread All 2.9562 2.8724 0.5081** Mixed quotes 2.4889 2.3869 0.5959** Eveneighth quotes 3.0754 3.1128 0.4178** Panel C: Realized Spread All 0.2874 0.3040+ 0.1280** Mixed quotes 0.1926 0.2252* 0.1394** Eveneighth quotes 0.3116 0.3431** 0.1162** Panel D: Percentage Realized Spread All 2.4074 2.3784+ 0.6280** Mixed quotes 2.0931 1.9967 0.7165** Eveneighth quotes 2.4876 2.5674 0.5369** **(*,+) indicates that the value is significantly different at 1% (5%, 10%) level from the preceding value in the same row.
-
28
Table VIII Effect of Listing Market, Odd-Eighth Quotes and OHR on IPO Underpricing over 2 periods, 1993 to 1996 and 1994 to 2005 (excluding 1997): Full Sample Table VIII reports the OLS regression estimates of IPO underpricing for the NYSE-listed IPOs and the NYSE-qualified NASDAQ IPOs for the periods of 1993 to 1996 and 1994 to 2005 excluding 1997. Model (1) examines the effect of odd-eighth usage on IPO underpricing and model (2) examines the effect of the OHR on IPO underpricing. The dependent variable for both models is the IPO initial return. To correct for heteroscedasticity, the heteroscedasticity consistent estimation method is used in the model. Full sample (1) (2) 1993-1996 1994-2005 NASDAQ -7.299 6.401 (1.28) (1.58) ODD*NASDAQ 14.980 (1.92)+ ODD-EIGHTH 1.438 (0.61) RULE*NASDAQ -10.332 (2.04)* RULE 1.459 (0.70) LN(MARKET VALUE) -1.529 -1.958 (1.23) (1.41) LN(AGE+1) -1.017 -0.443 (1.23) (0.69) INTERNET 0.294 2.426 (0.04) (0.43) HI-TECH -2.416 0.499 (0.84) (0.24) VC 0.956 0.160 (0.32) (0.07) UW RANKING -0.085 -1.068 (0.05) (0.86) LOCKUP DAYS -0.023 -0.056 (1.06) (3.01)** RETENTION 12.667 12.791 (2.62)* (3.50)** INTEGER PRICE -3.362 1.738 (1.56) (0.91) PRICE UPDATE 43.474 46.341 (4.00)** (5.46)** CONSTANT 20.631 33.092 (1.23) (2.30)* Observations 109 164 R-squared 0.45 0.47 Robust t statistics in parentheses + significant at 10%; * significant at 5%; ** significant at 1%
-
29
Table IX Effects of Odd-Eighth Quotes and Trading Systems on the Differences in IPO Underpricing within and across Markets ( the NYSE-listed IPOs vs. the NYSE-qualified NASDAQ-listed IPOs): Full Sample Table IX reports the test results of the effects of odd-eighth quotes, market listing and OHR within each market and across markets for IPOs issued in two periods. Panel A reports the effects of odd-eighth quotes for the period of 1993 to 1996. The Within Market difference in underpricing is based on post-1994 minus pre-1994. The Across Market difference in underpricing is based on NASDAQ minus NYSE. Panel B reports the effects of the OHR for the periods of 1994 to 2005 excluding 1997. The Within Market difference in underpricing is based on post-OHR minus pre-OHR. The Across Market difference in underpricing is based on NASDAQ minus NYSE. Full sample
Panel A: 1993-1994 vs. 1994-1996 (1) Within Market Null Hypothesis Sign p-value NYSE: (Post-1994) (Pre-1994) Odd=0 + 0.5442 NASDAQ: (Post-1994) (Pre-1994) Odd*NASDAQ +Odd=0 + 0.0275 (2) Across Market Null Hypothesis Sign p-value Pre-1994: NASDAQ NYSE NASDAQ=0 + 0.2022 Post-1994: NASDAQ - NYSE NASDAQ + Odd*NASDAQ=0 + 0.0801
Panel B: 1994-1996 vs. 1998-2005 (1) Within Market Null Hypothesis Sign p-value NYSE: (Post-OHR) (Pre-OHR) Rule=0 + 0.4840 NASDAQ: (Post-OHR) (Pre-OHR) Rule*NASDAQ + Rule=0 - 0.0532
(2) Across Market Null Hypothesis Sign p-value Pre-OHR: NASDAQ NYSE NASDAQ=0 + 0.1163 Post-OHR: NASDAQ - NYSE NASDAQ + Rule*NASDAQ=0 - 0.1615
-
30
Table X Effect of Listing Market, Odd-Eighth Quotes and OHR on IPO Underpricing over 2 periods, 1993 to 1996 and 1994 to 2005 (excluding 1997): Cold Issues Table X reports the OLS regression estimates of cold IPO underpricing for the NYSE-listed IPOs and the NYSE-qualified NASDAQ IPOs for the periods of 1993 to 1996 and 1994 to 2005 excluding 1997. Model (1) examines the effect of odd-eighth usage on IPO underpricing and model (2) examines the effect of the OHR on IPO underpricing. The dependent variable for both models is the IPO initial return. To correct for heteroscedasticity, the heteroscedasticity consistent estimation method is used in the model. (1) (2) 1993-1996 1994-2005 ODD-EIGHTH -0.057 (0.03) ODD*NASDAQ 9.118 (1.97)+ NASDAQ 9.347 (2.11)* RULE*NASDAQ -9.586 (1.86)+ RULE -0.542 (0.26) LN(MARKET VALUE) -0.396 0.750 (0.25) (0.71) LN(AGE+1) -0.578 -0.494 (0.92) (0.51) INTERNET 12.748 (5.37)** HI-TECH 2.164 1.237 (0.55) (0.58) VC -5.658 -4.140 (1.43) (1.27) UW RANKING -0.752 -0.152 (0.59) (0.11) LOCKUP DAYS -0.069 -0.021 (0.95) (0.87) RETENTION 0.946 1.890 (0.08) (0.35) INTEGER PRICE -0.367 0.877 (0.10) (0.50) PRICE UPDATE 10.830 8.640 (1.02) (1.07) CONSTANT 26.807 3.888 (1.18) (0.25) Observations 27 49 R-squared 0.55 0.40 Robust t statistics in parentheses + significant at 10%; * significant at 5%; ** significant at 1%
-
31
Table XI Effects of Odd-Eighth Quotes and Trading Systems on the Differences in IPO Underpricing within and across Markets ( the NYSE-listed IPOs vs. the NYSE-qualified NASDAQ-listed IPOs): Cold IPOs Table IX reports the test results of the effects of odd-eighth quotes, market listing and OHR within each market and across markets for cold IPOs issued in two periods. Panel A reports the effects of odd-eighth quotes for the period of 1993 to 1996. The Within Market difference in underpricing is based on post-1994 minus pre-1994. Panel B reports the effects of the OHR for the periods of 1994 to 2005 excluding 1997. The Within Market difference in underpricing is based on post-OHR minus pre-OHR. The Across Market difference in underpricing is based on NASDAQ minus NYSE.
Panel A: 1993-1994 vs. 1994-1996 Within Market Null Hypothesis Sign p-value NYSE: (Post-1994) (Pre-1994) Odd=0 - 0.9801 NASDAQ: (Post-1994) (Pre-1994) Odd*NASDAQ + Odd=0 + 0.1092
Panel B: 1994-1996 vs. 1997-2005 (1) Within Market Null Hypothesis Sign p-value NYSE: (Post-OHR) (Pre-OHR) Rule=0 - 0.7946 NASDAQ: (Post-OHR) (Pre-OHR) Rule*NASDAQ + Rule=0 - 0.0381
(2) Across Market Null Hypothesis Sign p-value Pre-OHR: NASDAQ NYSE NASDAQ=0 + 0.0417 Post-OHR: NASDAQ - NYSE NASDAQ + Rule*NASDAQ=0 - 0.9251
-
32
Appendix A Variable Definition
=
=
=
=
=
=
=
=
=+
=
=
=
Midpoint FilingMidpoint Filing - PriceOffer ePriceUpdat
otherwise 0 integer,an is priceoffer theif 1
gOutstandin Shares TotalOffered Shares - 1 Retention
shares their tradenot to agree insiders that IPO after the days ofnumber the Days Lockup
rankingquality ser' Underwrit RankingUW
otherwise 0 and capital, by venture backed is firm theif 1 VC
otherwise 0 and group,industry SDCtech -hi in the is firm theif 1 Tech -Hi
otherwise 0 firm,internet an as classified is firm theif 1 Internet
one, plus IPO at the age sfirm' theof logarithm natural the 1)ln(Age
issuanceafter goutstandin shares*priceoffer theas defined is which ue,market val sfirm' theof logarithm natural the Value)ln(Market
otherwise 0 and 1997year after issued is IPO theif 1 Rule
otherwise 0 and Nasdaqon listed is IPO theif 1 Nasdaq
Integer