# Factors that impact underpricing of Seasoned Equity Offerings.

The main goal of this article is to provide statistical evidence about factors that explain the level of Seasoned Equity Offering (SEO) underpricing. We study a sample of 1,840 Seasoned Equity Offerings (SEO) issued between 2003 and 2011. The most significant contribution of this article is the statistical evidence about a negative relationship between the offer price and SEO underpricing. This relationship might result from the conventional underwriter pricing practice. We report record levels of SEO underpricing never measured before, particularly in the year 2009 when 270 SEOs had an average level of underpricing of 6.86%. We also find that the level of SEO underpricing has dramatically increased since the 2008 financial crisis.Introduction

The main goal of this article is to analyze some factors that impact the level of Seasoned Equity Offering (SEO) underpricing calculated based on the pre-issue closing price. This is a phenomenon that has intrigued many financial professionals for decades, but no consensus exists about the specific factors that explain this type of underpricing. We analyze 1,840 SEOs issued between 2003 and 2001. Among the firms in our sample, 1,591 SEOs set an offer price different from the last closing price before the SEO date and only 249 firms used such a closing price as the relevant value for the share of stock.

This paper provides statistical evidence about the negative relationship between offer price and SEO underpricing. The lower the offer price, the higher the level of SEO underpricing. Also, we test some factors that impact the probability of underpricing a SEO. Although this is not the first paper to report a negative relationship between offer price and SEO underpricing, to the extent of our knowledge this paper is unique in that it provides statistical evidence on the significant increment of SEO underpricing after the financial crisis of 2008.

Some academic studies have reported and analyzed SEO underpricing. Smith (1977) compares two methods to raise additional equity capital: right offerings versus employing underwriters. He finds that regardless of the clear advantages of right offerings, most firms choose underwritten offerings. Smith studies 328 SEOs listed in the New York Stock Exchange (NYSE) and the American Stock Exchange (AMEX) from 1971 to 1975 and finds an average underprice of 0.54%. Parsons and Raviv (1985) propose a model of underwritten offerings of new SEOs where the offering price selected by the investment bank is lower than the market price. They argue that the underwriter must set an SEO price low enough to lure investors with high valuation of the issuer's new investment projects.

Bhagat and Frost (1986) study all primary public offerings of utility companies made from 1973 to 1980. They find an average underpricing level of -0.25%. Loderer, Sheehan, and Kadlec (1991) study 1,608 SEOs from 1980 to 1984. They find an average level of SEO underpricing of -1.41%. These results are driven by the level of SEO underpricing in the NASDAQ market where more than 80% of their SEOs are underpriced. They also find substantial variability in the SEO underpricing with ranges from -25 to +33%. Gerard and Nanda (1993) study the possibility of trading manipulation in SEOs. They argue that informed investors may attempt to influence offering prices by selling shares before the SEO, and profit subsequently from lower prices in the SEO.

Safieddine and Wilhelm (1996) study the nature and magnitude of short-selling activity around SEOs of 474 NYSE and AMEX listed firms In particular, they study the impact of the SEC adoption of the Rule 10b-21 to address concerns about manipulative short-selling practices. Using the close-to-offer return, they find an average SEO underpricing of -0.55% prior to the adoption of Rule 10b-21, and -0.46% after the adoption of such rule. Kim and Shin (2001) study a sample of 3,304 SEOs from 1983 to 1998. They find a mean of only 1.34% of SEO underpricing during the period before the adoption of Rule 10b-21 in 1988, but a mean of 2.99% after that year. Such a mean has been gradually increasing since 1988. They also find a more severe level of SEO underpricing in NASDAQ listed stocks. They conclude that Rule 10b-21 has raised discounts on SEO issues.

Bowen, Chen, and Cheng (2007) study 4,766 SEOs from 1984-2000 and find that more analyst coverage is associated with reduced SEO underpricing They find a mean underpricing of 2.38% in their sample. They find a SEO underpricing level of 4.05% for firms covered by one or zero analyst, to a level of 0.00% for firms covered by 31 or more analysts.

To the extent of our knowledge, the only previous articles that report a negative relationship between offer price and SE0 underpricing are Altinkilic and Hansen (2003) and Corwin (2003). Indeed, Altinkilic and Hansen (2003) find significant levels of SE0 discount for issues with low prices and for those with greater stock returns volatility. They also report larger discounts in offers by NASDAQ firms than those in NYSE/AMEX firms. They report an average level of SE0 underpricing of around 3% that has been stable throughout the 1990s.

Similarly, Corwin (2003) studies a sample of 6,637 SEOs between 1980 and 1998. She finds a positive relationship between SE0 underpricing and the level of uncertainty about the firm value, the relative offer size, the implementation of Rule 10b-21, and the concurrent level of underpricing in the IPO market. She also finds large underpricing for low-priced stocks and suggests that this phenomenon results from the conventional pricing practice of most underwriters.

The original contribution of this paper is that we report record levels of SE0 underpricing never measured before, particularly in the year 2009 when 270 SEOs had an average underpricing of 6.86%. The level of SE0 underpricing has increased dramatically after the financial crisis of 2008. We provide statistical evidence about the negative relationship between price and SEO underpricing. This relationship might result from conventional underwriter pricing practice.

Indeed, Lee, et al. (1996) suggest that SE0 underpricing might result from the underwriter's practice of pricing the offering at the bid price rounded down to the nearest eighth or integer. This notion is reinforced by Mola and Loughran (2004) who find evidence of significant clustering of offer prices at integers. We test the probability of underpricing an SE0 by using a logistic regression model. We find that this probability is negatively related with the SE0 offer price.

Also, our results suggest that underpricing is more likely to happen among firms listed in NASDAQ, among SEOs issued after the financial crisis of 2008, and in SEOs with high levels of stock price volatility.

The rest of this paper is organized as follows: section 1 describes our methodology and proposed models; section 2 explains our sample and provides some summary statistics; section 3 provides details of our empirical results; section 4 provides some interpretations of our findings, section 5 summarizes the major conclusions about our research work; and the last section lists the bibliographic references cited in this article.

The lower the offer price, the higher the level of SEO underpricing.

1. Methodology

The level of SEO underpricing (U[P.sub.1]) is measured by the 1-day holding period return for SEO i, as ([P.sub.1] - [P.sub.0])/[P.sub.0], where Po is the offer price and [P.sub.0], is the closing price of the day before the SEO date. The level of underpricing is determined for each firm in our sample of SEOs. We also use the difference-in-mean test (parametric) and the Mann-Whitney test (non-parametric) to analyze differences in the level of underpricing in our sample. The cross-sectional variations of the level of underpricing (U[P.sub.1]) are determined using the following ordinary least square (OLS) model:

[UP.sub.i] = [a.sub.1] + [b.sub.1] Ln[(Proc).sub.i] + [b.sub.2][Price.sub.i] + [b.sub.3]) [NASDAQ.sub.i] + [b.sub.4] [Supply.sub.i] + [b.sub.5][Crisis.sub.i] + [b.sub.6]U[R.sub.i] + [b.sub.7] S[D.sub.i] + [e.sub.i]

where [a.sub.1], is the intercept term of the model.

The variable Ln[(Proc).sub.i] is the natural logarithm of the total proceeds in U.S. dollars of SE0 i

Price, is the offer price of [SEO.sub.i]

[NASDAQ.sub.i] is a dummy variable that takes the value of one if SEO i is offered in the NASDAQ exchange and zero otherwise;

[Supply.sub.i] is the ratio of total shares offered in SEO i to total share outstanding after the offering;

[Crisis.sub.i] is a dummy variable that takes the value of one if SEO i is offered after 2008 and zero otherwise;

U[R.sub.i] is the underwriter reputation measured by the adjusted Carter-Manaster rankings from Jay Ritter's website at http://bear.cba.ufl.edu/ritter/rank.htm (see Carter and Manaster, 1990);

S[D.sub.i] is the standard deviation of daily returns in the 180 days prior to the offer; finally, e, is the error term.

We also apply a logistic regression model (probit mode!) to determine the factors that impact the probability of underpricing a SEO. Indeed, out of 1,840 SEOs in our sample, 249 SEOs did not experience any level of underpricing. The following probit model is considered:

In [[p.sub.t]/1 - [p.sub.t]] = [[alpha].sub.1] + [[beta].sub.1] Ln[(Proc).sub.i] + [[beta].sub.2] [Price.sub.i] + [[beta].sub.3][NASDAQ.sub.i] + [[beta].sub.4][Suppty.sub.i] + [[beta].sub.5] [Crises.sub.i] + [[beta].sub.6][UR.sub.i] + [[beta].sub.7][SD.sub.i] + [e.sub.i]

where [p.sub.i] is the probability that SEQ i will be underpriced and all variables and symbols at the right side of the equations are the same as those described above.

2. Sample and Summary Statistics

Our sample consists of 1,840 SEOs offered between 2003 and 2011; 402 firms have two consecutive SEOs; 160 firms have three consecutive SEOs; 68 firms have four consecutive SEOs, and 32 firms have five consecutive SEOs. The SEQ-related data is obtained from Securities Data Corporation's (SDC) Global New Issues database. The firms in our sample have stock price data available at the Center for Research in Security Prices (CRSP). The closing price of the day before the SEQ is obtained from CRSP. Those SEOs without price data at CRSP were eliminated. We excluded SEOs of American Depositary Receipts (ADRs), enhanced income securities, stocks with warrants, income depositary shares, unit SEOs, and foreign companies.

Data for our control variables is retrieved from the Securities and Exchange Commission (SEC) filings at the Electronic Data Gathering, Analysis, and Retrieval system (EDGAR) database. Following the same methodology of prior research, we use the daily volume as a reference to correct possible errors in the offer date. We apply this methodology when the trading volume is twice the average trading volume of the previous 250 days.

Exhibit 1 shows the average SEO underpricing per year. The average was increasing slightly between 2003 and 2008, but jumped to more than 6% after 2008. This record level of SEO underpricing has never been recorded before and constitutes evidence that the financial crisis of 2008 had a significant impact in the underwriting activity of securities firms. Exhibit 2 shows the results of T-tests to verify if the level of underpricing is statistically different from zero for the full sample and several sub-samples. Another important result is that there are not significant differences between the underpricing levels of consecutive SEOs.

Exhibit 1: SEO Underpricing per Year, Descriptive Statistics Year 2003 2004 2005 2006 2007 2008 2009 2010 SEO 2.96% 2.77% 2.81% 3.04% 3.22% 3.38% 6.86% 6.69% Undcrpricing #of SEOs 160 226 177 237 210 65 270 352 Std. Dev. 0.036 0.039 0.043 0.041 0.045 0.058 0.064 0.080 Year 2011 SEO 5.03% Undcrpricing #of SEOs 143 Std. Dev. 0.073 Level of Underpricing per Year 2003 3.0% 2004 2.8% 2005 2.8% 2006 3.0% 2007 3.2% 2008 3.4% 2009 6.9% 2010 6.7% 2011 5.0% Note: Table made from line graph. Exhibit 2: SEO Underpricing per Sub-sample N Mean T-test Skewness (Test value - 0) t-value Sic (2-talled) Full Sample 1839 4.42% 31.67 (0.000) 1.720 **** Sub-sample with two consecutive SEOs First SEO 402 4.45% 14.53 (0.000) 1.806 **** Second SEO 402 4.72% 14.80 (0.000) 1.6O0 **** Sub-sample with three consecutive SEOs First SEO 160 4.16% 8.55 (0.000) 1.855 **** Second SEO 160 4.03% 8.53 (0.000) 1.433 **** Third SEO 160 4.53% 8.49 (0.000) 1.594 **** Sub-sample with four consecutive SEOs First SEO 68 2.94% 5.430 (0.000) 2.209 **** Second SEO 68 3.51% 5.199 (0.000) 2.000 **** Third SEO 68 3.46% 5.072 (0.000) 2.345 **** Fourth SEO 68 3.55% 5.090 (0.000) 1.826 **** Sub-sample with Ave consecutive SEOs First SEO 32 2.10% 3.142 (0.003) *** 2.965 Second SEO 32 3.72% 3.280 (0.002) *** 1.852 Third SEO 32 2.77% 3.442 (0.001) *** 0.341 Fourth SEO 32 3.76% 2.877 (0.007) *** 1.566 Fifth SEO 32 4.48% 4.024 (0.000) *** 1.128 Kolmogorov-Smirnov Non-Parametric Tests One-Sample Test Full Sample 7.51 **** - Sub-sample with MWRST two consecutive SEOs First SEO 3.75 **** Z = -1.06 (0.287) Second SEO 3.37 **** Sub-sample with FPWA three consecutive SEOs First SEO 2.604 **** [Chi.sub.2]=0.33 (0.848) Second SEO 2.452 **** Third SEO 2.092 **** Sub-sample with FPWA four consecutive SEOs First SEO 1.861 *** [Chi.sub.2] = 1.68 (0.641) Second SEO 2.121 *** Third SEO L735 *** Fourth SEO 1.608 ** Sub-sample with FPWA Ave consecutive SEOs First SEO 1.637 *** [Chi.sub.2] = 6.88 (0.142) Second SEO 1.675 ** Third SEO 1.182 Fourth SEO 1.188 Fifth SEO 0.972 Notes: *, **, ***, and **** denote statistical significance at the 10%, 5%, 1%, and 0.1% significance levels, respectively. MWRST stands for Mann-Whitney Rank-Sum Test FOWA stands for riedman One-way ANOVA Test

Exhibit 3 provides summary statistics about our dependent and independent variables. The average level of SEO underpricing is 4.42%, which ranges between -10.26 and 33.18%. The average dollar amount raised in the SEOs is $182 million. The mean offer price was about $20 and the average ratio of shares offered was 18.78%.

Exhibit 3: SEO Underpricing, Descriptive Statistics Descriptive N Min. Max. Mean Std. Statistics Dev. SEO 1839 -10.26% 33.18% 4.42% 0.05991 Underpricing (%) Proceeds 1773 $0.462 $10,650 $182.46 467.776 (SMIL) Offer Price 1840 $0.1 $389.75 $20.24699 19.6968 ($) Percentage 1485 0.1718% 1,672.8% 18.78% 0.532728 Offered

Exhibit 4 shows the independent sample tests of two independent variables per exchange. The offer price was not included since statistical differences exist due to differences in the listing requirements. The only significant result is the average dollar amount raised in the SEO. Those firms that offer their shares in NASDAQ raised an average of $88 25 million, which is statistically different from the average of $293 7 million in those firms that listed their shares in other exchanges. There are no significant differences in the percentage of shares offered when comparing NASDAQ with other stock exchanges.

Exhibit 4: Independent Sample Tests of Some Independent Variables per Exchange Mean Levene's Test SEO Proceeds (Smil) from SEOs Offered in NASDAQ vs. Other Exchanges (N-960 vs. 813) NASDAQ $88.25 F= 106.71 (0.000) **** Other $293.7 Percentage of SEO Shares Offered in NASDAQ vs. Other Exchanges (N=811 vs. 674) NASDAQ 20.36% F = 0.4704 (0.493) Other 16.87% t-test for Equality of Means (Sig. 2-tailed) SEO Proceeds (Smil) from SEOs Offered in NASDAQ vs. Other Exchanges (N-960 vs. 813) NASDAQ Equal variances not t = -8.79 assumed (0.000) **** Other Percentage of SEO Shares Offered in NASDAQ vs. Other Exchanges (N=811 vs. 674) NASDAQ Equal variances t = 1.257 assumed (0.209) Other Notes: The p-values are shown In parentheses, *, **, *** and ****denote statistical significance at the 10%, 5%, 1%, and 0.1% significance levels, respectively.

3. Empirical Results

Exhibit 5 shows the independent sample tests of the level of SEO underpricing before and after the 2008 financial crisis. The average SEO underpricing before the 2008 financial crisis is 2.96%. This average is statistically different from the average underpricing after the 2008 crisis, which is 6.2%. These results are significant at 0.1% level of significance for both the parametric and non-parametric tests.

Exhibit 5: Independent Sample Tests of SEO Underpricing per Period. Mean Levenc's t-test for Test Equality of Means (Sic. 2-tailed) SEO Underpricing Before vs. After the 2008 Financial Crisis (N: 830 vs. 1009) Before 2009 6.20% F = 223.179 Equal (0.000) variances **** not assumed After 2008 2.96% Nonparamctric test (Mann.Whitney Test): Z-value = 11.8763 (0.000) **** SEO Underpricing Before vs. After the 2008 Financial Crisis (N: 830 vs. 1009) Before 2009 t= 11.37 (0.000) **** After 2008 Nonparamctric test (Mann.Whitney Test): Z-value = 11.8763 (0.000) **** Notes: The p-values are shown in parentheses, *, **, *** and **** denote statistical significance at the 10%, 5%, 1%, and 0.1% significance levels, respectively.

Exhibit 6 shows the independent sample tests of the level of SEO underpricing for SEOs listed on NASDAQ versus those listed in NYSE and AMEX. The average underpricing for firms listed in NASDAQ is 5.54%. This average is statistically different from the average underpricing in those firms listed in NYSE/AMEX which is 3.16%. These results are significant at a 0.1% significance level for both the parametric and non-parametric tests. These results are higher than those of Kim and Shin (2001) who find an average SEO underpricing of 2.21 to 4.35% from 1983 to 1998 for firms listed in NASDAQ. These results are also similar to those of Loderer, Sheehan, and Kadlec (1991) who find SEOs listed in NASDAQ more severely underpriced than those listed in NYSE and AMEX.

Exhibit 6: Independent Sample Tests of SEO Underpricing per Exchange. Mean Levene's t-test for Test Equality of Means (Sic. 2-tailed) NASDAQ vs.NYSE/AMEX ()980 vs. 895) NASDAQ 5.54% F= 85.679 Equal t = 8.87 (0.000) variances (0.000) **** not **** assumed Others 3.16% Non-parametric test (Mann-Whitney Test): Z-value = -9.133 (0.000) **** Notes: The p-values are shown in parentheses. *, **, ***, and **** denote statistical significance at the 10%, 5%, 1%, and 0.1% significance levels, respectively.

We ranked our sample based on the SEO offer price from lowest to highest. Then we compare the average SEO underpricing of the upper quintile versus that of the lower quintile, the upper quartile versus the lower quartile, and the first half of low-priced SEOs against the second half with high-priced offerings.

Exhibit 7 shows the independent sample tests of the average SEO underpricing for each group. The average level of underpricing of the first half of our sample with low offer prices is 6.45% while the second half with high offer prices has an average underpricing of 2.40%. Similarly, the average underpricing above and below the upper and lower quartiles are 8.60% versus 2.03%, and the upper and lower quintiles are 9.20% versus 1.94%.

Exhibit 7: Independent Sample Tests of SEO Underpricing per Offer Price. Mean Levene's t-test for t = Test Equality 13.999 of Means (0.000) (Sig. **** 2-tailed) 1st vs. 5th quintile (H: 367 vs. 368) 1st Quintile 9.20% F = Equal 390.726 variances not assumed 5th Quintile 1.94% Non-parametric test (Mann-Whitney Test): Z-value = -11.53 (0.000) **** 1st vs. 4th quintile (H: 459 vs. 460) 1st Quintile 8.60% F = Equal t = 428.789 variances 15.02 (0.000) not (0.000) **** assumed **** 4th Quintile 2.03% Non-parametric test (Mann-Whitney Test): Z-value = -12.78 (0.000) **** 1th half vs. 2nd half (N: 920 vs. 919) 1st Half 6.45% F = Equal t = 446.5064 variances 15.41 (0.000) not (0.000) **** assumed **** 2nd Half 2.40% Non-parametric test (Mann-Whitney Test): Z-value = -1379 (0.000) **** Notes: The p-values are shown in parentheses. *, **, ***, and **** denote statistical significance at the 10%, 5%, 1%, and 0.1% significance levels, respectively.

All these results are statistically significant at conventional levels of confidence for both the parametric and non-parametric tests. These results clearly suggest a positive relationship between the level of SEO underpricing and the offer price. This positive relationship is significant because the group (half, quartile, or quintile) with the lowest offer price exhibits the highest level of SEO underpricing, and the opposite is also true regarding the group with the highest offer price.

In order to discard the possibility that the results above are driven by the stock exchange in which the offering takes place, we repeat the same analysis, splitting our sample into two groups: firms listed in NASDAQ versus firms listed in other exchanges (NYSE and AMEX). We ranked each group based on the SEO offer price from lowest to highest. For each new group we compare the average SEO underpricing of the upper quintile versus that of the lower

Exhibit 8 shows the independent sample tests of the average SE0 underpricing of firms traded in NASDAQ. The average level of underpricing of the first half of this group (low offer prices) is 7.98%, while the second half (high offer prices) has an average underpricing of 3.10%. Similarly, the average level of SE0 underpricing above and below the upper and lower quartiles are 10.34% versus 2.57%, and the upper and lower quintiles are 10.68% versus 2.47%. All these results are statistically significant at conventional levels of confidence for both the parametric and non-parametric tests. These results suggest a positive relationship between the level of SEO underpricing and the offer price for firms listed in NASDAQ.

Exhibit 8: Independent Sample Tests of SEO Underpricing per Offer Price in NASDAQ Mean Levene's Test t-test for Equality of Means (Sig. 2-tailed) 1st vs. 5th quintile (N: 196 vs. 196) 1st Quintile 10.68% F= 136.72 Equal t= (0.000) **** variances -10.74(0.000) not **** assumed 5th Quintile 2.47% Non-parametric test (Mann-Whitney Test): Z-value = -8.0614 (0.000) **** 1st vs. 4th quintile (N: 245 vs. 245) 1st Quartile 10.34% P -266.268 Equal t = -11.76 (0.000) **** variances (0.000) **** not assumed 4th Quartile 2.57%1 Non-parametric test (Mann-Whitney Test): Z-value = -9.10355 (0.000) **** 1st half vs. 2nd half(N: 490 vs. 490) 1st Half 7.98% F= Equal t= -12.23 225.01(0.000) variances (0.000)**** **** not assumed 2nd Half 3.10% Non-parametric test (Mann-Whitney Test): Z-value = -10.5235 (0.000) **** Notes: The p-values are shown in parentheses. *, **, ***, and **** denote statistical significance at the 10%, 5%, 1%, and 0.1% significance levels, respectively.

Exhibit 9 shows the independent sample tests of the average SEO underpricing of firms traded in NYSE and AMEX. The average level of underpricing of the first half of this group (low offer prices) is 4.35% while the second half (high offer prices) has an average underpricing of 1.96%. Similarly, the average level of SEO underpricing above and below the upper and lower quartiles are 5.89% versus 1.72%, and the upper and lower quintiles are 6.56% versus 1.72%.

Exhibit 9: Independent Sample Tests of SEO Underpricing per Offer Price in NYSE/AMEX. Mean Levene's Test t-test for Equality of Means (Sig. 2-tailed) 1st vs. 5th quintile (N: 171 vs. 171) 1st Quintile 6.56% F= 95.22217 Equal t= (0.000) **** variances -8.043(0.000) not **** assumed 5th Quintile 1.72% Non-parametric test (Mann-Whitney Test): Z-value = -7.882 (0.000) **** 1 st vs. 4th quintile (N: 214 vs. 215) lst Quartile 5.89% P- 113.324 Equal t = -8.067 (0.000) **** variances (0.000) **** not assumed 4th Quartile 1.72% Non-parametric test (Mann-Whitney Test): Z-value = -7.7788 (0.000) **** 1st half vs. 2nd half(N: 429 vs. 430) 1st Half 4.35% F= Equal t= -7.597 11.50691(0.000) variances (0.000)**** **** not assumed 2nd Half 7.96% Non-parametric test (Mann-Whitney Test): Z-value = -7.3526 (0.000) **** Notes: The p-values are shown in parentheses. *, **, ***, and **** denote statistical significance at the 10%, 5%, 1%, and 0.1% significance levels, respectively.

All these results are statistically significant at conventional levels of confidence for both the parametric and non-parametric tests. These results suggest a positive relationship between the level of SEO underpricing and the offer price for firms traded in NYSE and AMEX.

In order to discard the possibility that the results above are driven by penny stocks, we repeat the same analysis splitting our sample into two groups: firms with an offer price below the median ($16.5) and firms with an offer price above the median. We ranked each group based on the SEO offer price from lowest to highest. For each new group we compare the average SEO underpricing of the upper quintile versus the lower quintile, the upper quartile versus the lower quartile, and the first half with low offer prices against the second half with high offer prices.

Exhibit 10 shows the independent sample tests of the average SEO underpricing of firms with an offer price less than or equal to $16.5 (median). The average level of underpricing of the first half of this group (low offer prices) is 8.60% while the second half (high offer prices) has an average underpricing level of 4.30%. Similarly, the average level of SEO underpricing above and below the upper and lower quartiles are 10.34% versus 3.6%, and the upper and lower quintiles are 10.33% versus 3.63%. All these results are statistically significant at conventional levels of confidence for both the parametric and non-parametric tests. These results confirm a positive relationship between the level of SEO underpricing and the offer price for firms priced at less than or equal to $16.5 (median).

Exhibit 10: Independent Sample Tests of SEO Underpricing per Offer Price for Firms with Offer Price Below the Median ($16.5) Mean Levene's t-test for Test Equality of Means (Sig. 2-tailed) 1 st vs. 5th quintile (N: 184 vs. 184) 1st Quintile 10.33% F= 213.797 Equal t= -7.626 (0.000) variances (0.000) **** **** not assumed 5th Quintile 3.63% Non-parametric test (Mann-Whitney Test): Z-value = -5.386 (0.000) **** 1 st vs. 4th quintile (N:203 vs. 230) lst Quartile 10.34% P -230.055 Equal t = -9.082 (0.000) variances (0.000) **** **** not assumed 4th Quartile 3.60%1 Non-parametric test (Mann-Whitney Test): Z-value = (0.000) **** 1st half vs. 2nd half(N: 460 vs. 460) 1st Half 8.60% F= 201.288 Equal t= -9.226 (0.000) variances (0.000)**** **** not assumed 2nd Half 4.30% Non-parametric test (Mann-Whitney Test): Z-value = -7.7378 (0.000) **** Notes: The p-values are shown in parentheses. *, **, ***, and **** denote statistical significance at the 10%, 5%, 1%, and 0.1% significance levels, respectively.

Exhibit 11 shows the independent sample tests of the average SEO underpricing of firms with an offer price greater than $16.5 (median). The average level of underpricing of the first half of this group (low offer prices) is 2.76%, while the second half (high offer prices) has an average underpricing of 2.03%. Similarly, the average level of SEO underpricing above and below the upper and lower quartiles are 2.94% versus 1.90%, and the upper and lower quintiles are 3.28% versus 1.88%. All these results are statistically significant at conventional levels of confidence for both the parametric and non-parametric tests. These results confirm a positive relationship between the level of SEO underpricing and the offer price for firms with an offer price greater than $16.5 (median).

Exhibit 11: Independent Sample Tests of SEO Underpricing per Offer Price for Firms with Offer Price Above the Median ($16.5) Mean Levene's t-test for Test Equality of Means (Sig. 2-tailed) 1 st vs. 5th quintile (N: 184 vs.183) 1st Quintile 3.28% F= 6.4352 Equal variances t= (0.000) not assumed -3.67(0.000) **** **** 5th Quintile 1.88% Non-parametric test (Mann-Whitney Test): Z-value = -4.393 (0.000) **** 1 st vs. 4th quintile (N: 230 vs. 229) lst Quartile 2.94% F = Equal variances t = -3.697 -7.9418 not assumed (0.000) (0.000) **** **** 4th Quartile 1.90% Non-parametric test (Mann-Whitney Test): Z-value = -4.3081 (0.000) **** 1st half vs. 2nd half(N: 920 vs. 920) 1st Half 2.76% F= 7.3068 Equal variances t= -3.726 (0.000) not assumed (0.000)**** **** 2nd Half 2.03% Non-parametric test (Mann-Whitney Test): Z-value = -4.383 (0.000) **** Notes: The p-values are shown in parentheses. *, **, ***, and **** denote statistical significance at the 10%, 5%, 1%, and 0.1% significance levels, respectively.

In other to discard the possibility that the results above are driven by the impact of the financial crisis of 2008, we repeat the same analysis splitting our sample into two groups: SEOs issued during or before 2008 and SEOs issued after 2008. We ranked each group based on the SEO offer price from lowest to highest. For each new group we compare the average SEO underpricing of the upper quintile versus the lower quintile, the upper quartile versus the lower quartile, and the first half with low offer prices against the second half with high offer prices.

Exhibit 12 shows the independent sample tests of the average underpricing of SEOs that took place during or before 2008. The average underpricing of the first half of this group (low offer prices) is 4.21%, while the second half (high offer prices) has an average underpricing of 1.76%. Similarly, the average level of SEO underpricing above and below the upper and lower quartiles are 5.71% versus 1.51%, and the upper and lower quintiles are 6.11% versus 1.53%. All these results are statistically significant at conventional levels of confidence for both the parametric and non-parametric tests. These results confirm a positive relationship between the level of SEO underpricing and the offer price for offerings that took place during or before 2008.

Exhibit 12: Independent Sample Tests of SEO Underpricing per Offer Price for Offerings Before 2009 Mean Levene's t-test for Test Equality of Means (Sig. 2-tailed) 1 st vs. 5th quintile (N: 215 vs. 215) 1st Quintile 6.11% F= 115.72 Equal t= 9.93 (0.000) variances not (0.000) **** **** assumed 5th Quintile 1.53% Non-parametric test (Mann-Whitney Test): Z-value = -9.208 (0.000) **** 1 st vs. 4th quintile (N: 268 vs. 269) lst Quartile 5.71% F =134.3124 Equal t = 10.502 (0.000) variances not (0.000) **** **** assumed 4th Quartile 1.51% Non-parametric test (Mann-Whitney Test): Z-value = -9.88 (0.000) **** 1st half vs. 2nd half(N: 537 vs. 537) 1st Half 4.21% F= 118.844 Equal t= 9.94 (0.000) variances not (0.000)**** **** assumed 2nd Half 1.76% Non-parametric test (Mann-Whitney Test): Z-value = -9.6409 (0.000) **** Notes: The p-values are shown in parentheses. *, **, ***, and **** denote statistical significance at the 10%, 5%, 1%, and 0.1% significance levels, respectively.

Exhibit 13 shows the independent sample tests of the average underpricing of SEOs that took place after 2008. The average level of underpricing of the first half of this group (low offer prices) is 8.59%, while the second half (high offer prices) has an average underpricing of 4.3%. Similarly, the average level of SEO underpricing above and below the upper and lower quartiles are 10.46% versus 3.6%, and the upper and lower quintiles are 10.24% versus 3.52%. All these results are statistically significant at conventional levels of confidence for both the parametric and non-parametric tests. These results confirm a positive relationship between the level of SEO underpricing and the offer price for offerings that took place after 2008.

Exhibit 13: Independent Sample Tests of SEO Underpricing per Offer Price for Offerings After 2008 Mean Levene's t-test for Test Equality of Means (Sig. 2-tailed) 1 st vs. 5th quintile (N: 153 vs. 153) 1st Quintile 10.24% F= 277.603 Equal t= -6.91 (0.000) variances not (0.000) **** **** assumed 5th Quintile 3.52% Non-parametric test (Mann-Whitney Test): Z-value = -4.5113 (0.000) **** 1 st vs. 4th quintile (N: 192 vs. 192) lst Quartile 10.46% F = 272.536 Equal t = -8.30 (0.000) variances not (0.000) **** **** assumed 4th Quartile 3.60% Non-parametric test (Mann-Whitney Test): Z-value = -6.1999 (0.000) **** 1st half vs. 2nd half(N:382 vs. 382) 1st Half 8.59% F= 223.66 Equal t= -8.397 (0.000) variances not (0.000)**** **** assumed 2nd Half 4.30% Non-parametric test (Mann-Whitney Test): Z-value = -6.888 (0.000) **** Notes: The p-values are shown in parentheses.*, **, ***, and **** denote statistical significance at the 10%, 5%, 1%, and 0.1% significance levels, respectively.

We also perform a panel study on the level of SEO underpricing using our full sample and several sub-samples. Exhibit 14 shows the results of five linear regression models. The first model tests our full sample and includes several control variables mentioned in previous academic articles. The results confirm again that the level of SEO underpricing has a significant and negative relationship with the offer price ([Price.sub.i]): the higher the offer price, the lower the level of SEO underpricing. This result is significant in all models and consistent with those of table 7.

Exhibit 14: Regression Model with Panel Data on the Level of SEO Underpricing Model 1 Model 2 (After (General) 2008) Intercept 0.132436 (4.236) 0.118377 (2.759) **** *** Ln[(Proc).sub.i] -0.005092 -0.002882 (-1.21) (-2.91) *** [Price.sub.i] -0.000621 -0.001074 (-3.949)**** (-6.41)**** [NASDAQ.sub.i] 0.011766 0.014258 (3.536)**** [Supply.sub.i] -0.0000377 0.006993 (119) (-0.026827) [Crisis.sub.i] 0.023058 - (8.502)**** [SD.sub.i] 0.001724 0.001559 (1.91) * (3.33)**** [UR.sub.i] -0.001101 -0.000728 (-1.75) **** (-0.837936) [R.sup.2] 0.166627 0.09034 Adjusted 0.162458 0.08236 [R.sup.2] F-statistic 39.96014 11.32154 N 1407 691 Model 3 (Before Model 4 (NASDAQ) Model5 2009) (NYSEAMEX) Intercept 0.162827 0.199561 (5.666) 0.076141 (5.7268) **** **** (2.90825) *** Ln[(Proc).sub.i] -0.006661 -0.008221 (-3.9) -0.002103 (-4.43)**** **** (-1.340335) [Price.sub.i] -0.000385 -0.000669 -0.000683 (-2.955)*** (-3.44)**** (-4.938)**** [NASDAQ.sub.i] 0.009262 - - (3.344)**** [Supply.sub.i] -0.001684 -0.001311 0.007204 (-3.85)**** (-0.404613) (1.308106) [Crisis.sub.i] 0.023024 0.021102 (4.865)**** (5.832503)**** [SD.sub.i] 0.001311 0.002634 (2.344) 0.001123 (1.62) (2.645) *** ** [UR.sub.i] -0.001683 -0.001372 (-1.69) -0.000616 (-235)** (-0.849269) [R.sup.2] 0.159021 0.137771 0.161368 Adjusted 0.154413 0.131035 0.153317 [R.sup.2] F-statistic 34.50898 20.4525 20.04348 N 735 775 632 Notes: The t-values are shown in parentheses. *, **, ***, and ****denote statistical significance at the 10%, 5%, 1%, and 0.1% significance levels, respectively. The t-statistics are computed using the Newey-West (1987) covariance estimators. The t-statistics are also estimated using the White (1980) method with identical results in terms of significance, but they are not included in this article. The independent variables have high levels of tolerance.

The dummy variable [NASDAQ.sub.i] has a significant and positive relationship with the level of SEO underpricing. This result is consistent with those of Exhibit 6 and also consistent with the findings of Altinkilic and Hansen (2003), Corwing (2003), and Mola and Loughran (2004). The size of the SEO measured by the natural logarithm of the offering proceeds (Ln[(Proc).sub.i]) has a negative and significant relationship with the level of SEO underpricing. The relative SEO size (Supply,) is not significant at conventional levels of confidence. This result is consistent with those of Kim and Shin (2001), who find no significant relationship between relative offer size and SEO underpricing.

The dummy variable to control for the effect of the financial crisis of 2008 ([Crisis.sub.i]) has a positive and significant relationship with the level of SEO underpricing. This last result confirms those of Exhibit 1 and 5 that show a significant increase in the level of SEO underpricing after 2008. The volatility of the stock price ([SD.sub.i]) has a positive a significant relationship with the level of SEC) underpricing. This result is consistent with those reported in prior research (see Bowen, Chen, and Cheng, 2007). Finally, the underwriter reputation ([UR.sub.i]) has a negative relationship with the level of SEO underpricing, which is consistent with the findings of Mola and Loughran (2004).

The second model tests those SEOs issued after 2008. The results are the same as those using our full sample, except that the size of the SEO (Ln[(Proc).sub.i]) and the underwriter reputation ([UR.sub.i]) have no significant impact on the level of SEO underpricing.

The third model tests those SEOs issued before 2009, and the results are the same as those of our full sample. The fourth model tests those SEOs listed in NASDAQ, only, and the results are identical to those using our full sample. Finally, the fifth and last model tests those SEOs listed in NYSE and AMEX. In this last model the only significant results are the offer price ([Price.sub.i]) and the dummy variable to control for the effect of the financial crisis of 2008 ([Crisis.sub.i]) .

We also test whether the control variables of our panel study have explanatory power on the probability of underpricing a SEO. We test five logistic (probit) regression models. The results provide evidence that the probability of SEO underpricing is negatively related with the offer price ([Price.sub.i]): the higher the offer price, the lower the probability of SEO underpricing. Also, the results suggest that underpricing is more likely to happen among firms listed in NASDAQ, among SEOs issued after the financial crisis of 2008, and in SEOs with high levels of stock price volatility ([SD.sub.i]). These control variables are statistically significant at conventional levels of confidence. The second model tests those SEOs issued after 2008.

The results show that after 2008 no control variable has explanatory power on the probability of SEO underpricing, except for the ratio of total shares offered to total share outstanding after the offering ([Supply.sub.i]). This result suggests that the higher the proportion of shares offered, the higher the probability of underpricing the SEO. The third model tests those SEOs issued before 2009, and the results are the same as those obtained using the full sample.

The fourth model tests those SEOs listed in NASDAQ, only, and the results are the same as those obtained using our full sample, except for the offer price (Price) that has no significant influence in the probability of SE0 underpricing.

Finally, the fifth and last model tests those SEOs listed in NYSE and AMEX. The results are the same as those using our full sample, except for the underwriter reputation (UR) that has a significant negative impact on the probability of SEO underpricing.

Exhibit 15: Probit Regression Model of the Probability of SEO Underpricing Model 1 (General) Model 2 (After 2008) Intercept 0.071746 (0.0927) 0390292 (0.341593) Ln[(Proc).sub.i] 0.046852 (1.0405) 0.070639 (1.0054) [Price.sub.i] -0.01445 (-3.52) -0.009156 **** (-1.03206) [NASDAQ.sub.i] 0373825 (3.65) -0.02477 (-0.145) **** [Supply.sub.i] -0.01951 (-0.267) 1.768224 (2.0396) ** [Crisis.sub.i] 0.677991 (6.44) - **** [SD.sub.i] 0.10248 (3.51) 0.08061 i (1.4077) **** UR, -0.016473 (-0.84) -0,041007 (-1315796) McFadden [R.sup.2] 0.101468 0.030786 LR statistic 101.1398 8.162754 N 1407 691 No. Underpricing 160 33 Model 3 (Before Model 4 (NASDAQ) 2009) Intercept 1.336847 1.825944 (1.250939) (1.531487) Ln[(Proc).sub.i] -0.029163 -0.036394 (-0.487756) (-0.523423) [Price.sub.i] -0.013789 -0.008738 (-2.880) *** (-1.630483) [NASDAQ.sub.i] 0.528651 (4.105) - **** [Supply.sub.i] -0.05111 -0.053825 (-0.74721) (-0.7759) [Crisis.sub.i] - 036052 (2.531) ** [SD.sub.i] 0.098717 (2.797) 0.070305 (1.7951) *** * UR, -0.007379 0.012634 (-0.278711) (0.527711) McFadden [R.sup.2] 0.069205 0.028366 LR statistic 46.31862 11.83655 N 716 775 No. Underpricing 127 59 Model 5 (NYSE/AMEX) Intercept -0.777762 (-0.753515) Ln[(Proc).sub.i] 0.097605 (1.56414) [Price.sub.i] -0.016406 (-2.544334) ** [NASDAQ.sub.i] - [Supply.sub.i] 1.058972 (1.310676) [Crisis.sub.i] 0.964792 (6303) **** [SD.sub.i] 0.11254 (2.5678) ** UR, -0.059316 (-1.807) * McFadden [R.sup.2] 0.148823 LR statistic 82.64819 N 632 No. Underpricing 101 Notes; The Z-statistics are shown in parentheses.*, **, ***, and **** denote statistical significance at 10%, 5%, 1%, and 0.1% significance levels, respectively. Model 1 includes shelf registrations while model 2 excludes them. We found no multicouuiearity problems with Allison's [1999) methodology by estimating the equivalent linear regression model and evaluating the tolerance and the variance inflation factor for each independent variable. The standard error was estimated by using the generalized linear model method. The Z-statistics are estimated using the quasi-maximum likelihood (Huber/White) method.

4. Interpretations and implications

The results above provide statistical evidence that SE0 underpricing is a phenomenon driven primarily by the offer price. There is a negative and significant relationship between the offer price of an SE0 and its probability and magnitude of underpricing. The lower the offer price, the higher the level of SE0 underpricing. The control variables also confirm this negative relationship. The higher level of SEO underpricing for firms listed in NASDAQ versus that of firms listed in NYSE and AMEX also suggest this relationship. Firms listed in NASDAQ have an average offer price lower than those listed in NYSE and AMEX, and this might explain the difference in SEO underpricing. Exhibit 16 provides evidence that the average offer price of firms listed in NASDAQ is ($16.77) significantly lower than that of firms listed in NYSE and AMEX ($24.22).

Exhibit 16: Independent Sample Tests of Average SEO Offer Price per Exchange Mean Levene's t-test for Test Equality of Means (Sig. 2-tailed) NASDAQ vs. NYSE/AMEX (981 vs. 859) NASDAQ $16.77 F= 16.664 Equal t= -7.45 (0.000) variances (0.000) **** not **** assumed Others $24.22 Non-parametric test (Mann-Whitney Test): Z-value = -10.5 (0.000) **** Notes: The p-values are shown in parentheses. *, **, **, and ****, denote statistical significance at the 10%, 5%, 1%, and 0.1% sIgnificance levels, respectively.

Similarly, the significant difference in SEO underpricing between issues offered before and after the financial crisis of 2008 is also driven by offer prices. The negative impact of the financial crisis on the stock market is well known. In 2008, the Standard and Poor's 500 stock index, a broad barometer of the stock market, fell 37.2%. Not surprisingly, the average offer price of SEOs issued after 2008 was lower than the average for SEOs issued during or before that year. Correspondingly, the average SEO underpricing peaked in year 2009 and continued at high levels in 2010 and 2011. The chart below shows the average level of underpricing per year for firms listed in NASDAQ and those listed in NYSE and AMEX. The pattern is similar for both sub-samples.

Level of Underpricing per Year and Exchange NASDAQ NYSE & AMEX 2003 4.71% 1.63% 2004 4.18% 2.36% 2005 4.36% 2.19% 2006 4.08% 2.60% 2007 3.57% 3.06% 2008 2.96% 3.09% 2009 7.41% 4.56% 2010 7.81% 4.24% 2011 6.59% 3.54% Note: Table made from line graph.

Similarly, the size of the issue measured by the natural logarithm of the offering's proceeds is also related with the offer price. Large issues usually have relatively high offer prices, and penny stocks commonly need to sell millions of shares to raise modest amounts of capital. The negative relationship between the natural logarithm of the proceeds and the level of underpricing also reinforce this notion that SE0 underpricing is driven primarily by the offer price.

To test this interpretation, we ranked our sample based on the SEO offer price from lowest to highest. Then, we compared the average SEG proceeds of the upper quintile versus that of the lower quintile, the upper quartile versus the lower quartile, and the first half of low-priced SEOs against the second half with high-priced offerings. Exhibit 17 shows the independent sample tests of the average SE0 proceeds for each group. The average dollar proceeds of SEOs with low offer prices have associated lower proceeds than SEOs with high offer prices. All these results are statistically significant at conventional levels of confidence for both the parametric and non-parametric tests.

Exhibit 17: Independent Sample Tests of SEP Proceeds per offer Price Mean Levene's t-test for Test Equality of Means (Sig. 2-tailed) 1 st vs. 5th quintile (N: 369 vs. 365) 1st Quintile $42.8M F= 125.74 Equal (0.000) variances **** not assumed 5th Quintile $373M Non-parametric test (Mann-Whitney Test): Z-value = -16.2 (0.000) **** 1 st vs. 4th quintile (N: 461 vs. 461) lst Quartile $49M F - Equal 139.95 variances (0.000) not **** assumed 4th Quartile $286M Non-parametric test (Mann-Whitney Test): Z-value = -17.54 (0.000) **** 1st half vs. 2nd half(N: 904 vs. 869) 1st Half $82.7M F= 95.03 Equal (0.000) variances **** not assumed 2nd Half $286M Non-parametric test (Mann-Whitney Test): Z-value = -19.85 (0.000) **** 1 st vs. 5th quintile (N: 369 vs. 365) 1st Quintile t= -10.04(0.000) **** 5th Quintile Non-parametric test (Mann-Whitney Test): Z-value = -16.2 (0.000) **** 1 st vs. 4th quintile (N: 461 vs. 461) lst Quartile t = -10.76 (0.000) **** 4th Quartile Non-parametric test (Mann-Whitney Test): Z-value = -17.54 (0.000) **** 1st half vs. 2nd half(N: 904 vs. 869) 1st Half t= -9.21 (0.000) **** 2nd Half Non-parametric test (Mann-Whitney Test): Z-value = -19.85 (0.000) **** Notes: The p-values are shown in parentheses., *, **, ***, and **** denate statistical significance at the 10%, 5%, 1%, and 0.1% significance levels, respectively.

Correspondingly, the underwriter reputation measured by the adjusted Carter-Manaster rankings is also related with the SEO's offer price. SEO of small companies and penny stocks are rarely underwritten by prestigious investment banks. In the same way, underwriters with high reputations will prefer to underwrite large offerings to maximize the underwriting fees. The negative relationship between the underwriter reputation and the level of underpricing also reinforces this notion that SEO underpricing is driven primarily by the offer price.

To test this interpretation, we ranked our sample based on the SEO offer price from lowest to highest. Then, we compare the average underwriter reputation of the upper quintile versus that of the lower quintile, the upper quartile versus the lower quartile, and the first half of low-priced SEOs against the second half with high-priced offerings. Exhibit 18 shows the independent sample tests of the average underwriter reputation for each group The average underwriter reputation of SEOs with low offer prices is lower than that of SEOs with high offer prices. All these results are statistically significant at conventional levels of confidence for both the parametric and non-parametric tests.

Corwing (2003) suggests that the rounded offer prices suggested by Lee et. al (1996) results from the inaccurate nature of the SEO pricing process. She argues that offer price rounding implies a negative relationship between SEO underpricing and offer price since the rounded portion of the price represents a larger fraction of a lower price. She provides evidence that SEOs of low-priced securities are more underpriced than offers of high-priced securities. If price is a critical factor in SEO underpricing, additional analyses based on offer price would contribute to validating the most accepted hypothesis about SEO underpricing.

Exhibit 18: Independent Sample Tests of Underwriter Reputation per Offer Price Mean Levene's Test t-test for Equality of Means (Sig. 2-tailed) 1st vs. 5th quintile (N: 369 vs. 365) 1st Quintile 4.08 F= 136.72 (0.000) Equal **** variances not assumed 5th Quintile 7.54 Non-parametric test (Mann-Whitney Test): Z-value = -13.87 (0.000) **** 1st vs. 4th quintile (N: 461 vs. 461) lst Quartile 4.48 P- 192.44 (0.000) Equal **** variances not assumed 4th Quartile 7.51 Non-parametric test (Mann-Whitney Test): Z-value = -13.73 (0.000) **** 1st half vs. 2nd half(N: 920 vs. 920) 1st Half 5.7 F= 225.01(0.000) Equal **** variances not assumed 2nd Half 7.3.8 Non-parametric test (Mann-Whitney Test): Z-value = -11.5 (0.000) **** 1st vs. 5th quintile (N: 369 vs. 365) 1st Quintile t= -15.45(0.000) **** 5th Quintile Non-parametric test (Mann-Whitney Test): Z-value = -13.87 (0.000) **** 1st vs. 4th quintile (N: 461 vs. 461) lst Quartile t = -14.9 (0.000) **** 4th Quartile Non-parametric test (Mann-Whitney Test): Z-value = -13.73 (0.000) **** 1st half vs. 2nd half(N: 920 vs. 920) 1st Half t= -12.06 (0.000)**** 2nd Half Non-parametric test (Mann-Whitney Test): Z-value = -11.5 (0.000) **** Notes. The p-values are shown in parentheses.*, **, ***, and **** denote statistical significance at the 10%, 5%, 1%, and 0.1% significance levels, respectively.

5. Conclusions

The most significant contribution of this article is the statistical evidence about the negative relationship between the offer price and the level of SEO underpricing. This relationship might result from conventional underwriter pricing practice. We report record levels of SEO underpricing never recorded before, particularly in the year 2009 when 270 SEOs had an average underpricing of 6.86%. We find that the level of SEO underpricing has dramatically increased since the 2008 financial crisis. We also provide statistical evidence that the probability of SEO underpricing is negatively related with the offer price. Also, our results suggest that underpricing is more likely to happen among firms listed in NASDAQ, among SEOs issued after the financial crisis of 2008, and in SEOs with high levels of stock price volatility.

Winner of the Best Paper Award at the 6th Annual Symposium of the Financial Services Institute, International Dimensions of New Regulations: Effects on Consumers, Corporate Governance, Financial Markets and Accounting Practice. field at St John's University, 101 Murray Street, New York City, September 8-10, 2011 .

References

Allison, R D. (1999). Logistic Regression Using SAS[R] System: Theory and Application. Cary, NC: SAS Institute Inc.

Altinkilic, 0. and Hansen, R. S. (2003). Discounting and underpricing in seasoned equity offers. Journal of Financial Economics, 69(2), 285-323.

Bhagat, S. and Frost, R A. (1986). Issuing costs to existing shareholders in competitive and negotiated underwritten public utility equity offerings. Journal of Financial Economics, 15(1-2), 233-259.

Bowen, R. M.; Chen, X.; and Cheng, Q. (2007). Analyst coverage and the cost of raising equity capital: Evidence from underpricing of seasoned equity offerings. Working paper, University of Washington Business School and University of British Columbia.

Carter, Richard, and Steven Manaster. (1990). Initial public offerings and underwriter reputation. Journal of Finance, Vol. 45, pp. 1045-1067.

Corwin, S. A. (2003). The determinants of underpricing for seasoned equity offers. Journal of Finance, 58(5), 2249-79.

Gerard, B. and Nanda, V. (1993). Trading and manipulation around seasoned equity offerings. Journal of Finance, 48(1), 213-245.

Kim, K. A. and Shin, H. H. (2001). The underpricing of seasoned equity offerings: 1983-1998. Working paper. University of Wisconsin-Milwaukee and State University of New York.

Loderer, C. F.; Sheehan, D. and Kadlec, G. B. (1991). The Pricing of Equity Offerings. Journal of Financial Economics, 29(1), 34-58.

Lee, I., Lochhead, S., Ritter, J. and Zhao, Q. (1996). The costs of raising capital. Journal of Financial Research, 19(1), 59-74.

Mola, S., and Loughran, T. (2004). Discounting and clustering in seasoned equity offering prices. Journal of Financial and Quantitative Analysis, 39(1), 1-24.

Newey, W. and West, K. (1987). A simple Positive Semi-Definite, Heteroskedasticity and Autocorrelation Consistent Covariance Matrix. Econometrica, 55,703-708.

Parsons, J. and Raviv, A. (1985). Underpricing of seasoned issues. Journal of Financial Economics, 14(3), 377-397.

Safieddine, A. and Wilhelm Jr., W. J. (1996). An Empirical Investigation of Short-Selling Activity Prior to Seasoned Equity Offerings. Journal of Finance, 51,729-749.

Smith Jr., Clifford W. (1977). Alternative methods for raising capital: Rights versus underwritten offerings. Journal of Financial Economics, 5(3), 273-307.

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

Dr. Juan M. Dempere, Metropolitan State College of Denver, Colorado jdempere@mscd.edu

Printer friendly Cite/link Email Feedback | |

Author: | Dempere, Juan M. |
---|---|

Publication: | Review of Business |

Article Type: | Column |

Geographic Code: | 1USA |

Date: | Dec 22, 2012 |

Words: | 8877 |

Previous Article: | The duality of social enterprise: a framework for social action. |

Next Article: | Five necessary attitudes of a servant leader. |

Topics: |