 

 

[IDnAhv

Michigan State
University

 

 

 

This is to certify that the
dissertation entitled

THREE ESSAYS IN CORPORATE FINANCE

presented by

Hoontaek Seo

has been accepted towards fulfillment
of the requirements for the

Ph. D. degree in Finance

 

 

m J. Wu

Major Professor's Signature
(In: c «Q $9 20 0?

Date

 

MSU is an Afﬁrmative Action/Equal Opportunity Employer

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:IProi/Aoc&Pres/ClRC/DateDue.indd

THREE ESSAYS IN CORPORATE FINANCE
By

Hoontaek $60

A DISSERTATION

Submitted to
Michigan State University
In partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Finance

2009

ABSTRACT
THREE ESSAYS IN CORPORATE FINANCE
By

Hoontaek Seo

This dissertation examines the role of corporate governance in corporate
ﬁnancing decisions. The ﬁrst essay explore the effect of misalignment of
managerial interests with those of shareholders, which is measured by divergence
between managerial voting and cash ﬂow rights on the seasoned equity offerings
(SEOS) by dual-class ﬁrms. We Show that SEO announcement returns and long-run
stock performance following SEOS are negatively related to measures of the
divergence between managerial control and cash ﬂow rights. Our results support
the view of agency problems that the misalignment of interests between managers
and shareholders can create managerial incentives to undertake value-destroying
investments to extract private beneﬁts, ultimately leading to a reduction in ﬁrm
value.

The second essay investigates the effect of classiﬁed boards on the market
reaction to seasoned equity offering (SEO) announcements and the operating
performance following SEOS. We ﬁnd that ﬁrms with classiﬁed boards on average
earn lower SEO announcement returns relative to ﬁrms with unitary boards. Result
from the change in matched-ﬁrm-adjusted operating performance analysis shows
that ﬁrms having a classiﬁed board structure earn worse abnormal operating

performance following SEOS relative to ﬁrms having a unitary board structure. Our

results support the view that classiﬁed boards entrench managers and are
ineffective in preventing them from misuse of funds raised in SEOS.

The third essay examines the effect of product market competition on ﬁrms“
credit ratings. We ﬁnd that market competition is positively related to shareholder
value and negatively related to ﬁrm’s credit rating. We also ﬁnd that firms are more
likely to receive investment-grade credit ratings in less competitive markets.
Finally, we ﬁnd that ﬁrms in more competitive markets tend to make more R&D
and advertising expenditures. Overall, our results indicate that product market
competition align more closely the interests of managers and shareholders and thus
create managerial incentives to undertake riskier investments to maximize

shareholder value, which lowers ﬁrms’ credit ratings.

T 0 my wife, Bongsoon, for her unwavering love and support.

iv

ACKNOWLEDGMENTS

I would like to express my appreciation to the members of my dissertation
committee: Dr. Charles Hadlock, Dr. G. Geoffrey Booth, Dr. Ted Fee, and Dr.
Michael Mazzeo, for their helpful comments and guidance. I would particularly
like to express my deepest appreciation to my adviser, Dr. Charles Hadlock for his
help and assistance throughout the entire term of my doctoral program.

I am deeply indebted to Dr. G. Geoffrey Booth for supporting me in every
way possible throughout my doctoral program. I thank the other professors of the
ﬁnance department for sharing their knowledge and providing invaluable support
system. I extend many thanks to my fellow doctoral students in the ﬁnance
department.

My deepest gratitude goes to my family in Korea for their love and
continued support. I would like to thank my wife, Bongsoon and my children,
Joonho and Ji-In for their patience and understanding of the time I spent away from

them while completing this work.

TABLE OF CONTENTS

List of Tables ......................................................................................................... viii

1 Agency problems at dual-class ﬁrms: Evidence from seasoned equity

offerings
1.1 Introduction ......................................................................... 1
1.2 Sample Construction ............................................................... 4
1.3 Analysis of SEO announcement returns ......................................... 6
1.3.1 Divergence between voting rights and cash ﬂow rights. . . . . . . . .....6
1.3.2 Variable deﬁnition ........................................................ 7
1.3.3 Univariate analysis ....................................................... 10
1.3.4 Multivariate analysis ..................................................... 12
1.4 Analysis of long-run stock performance following SEOs ................... 14
1.4.1 Long-run abnormal returns ............................................. 14
1.4.2 Factor regressions ......................................................... 18
1.5 Conclusion ......................................................................... 19
Chapter 1 Appendix .......................................................................... 21
2 Classiﬁed boards and managerial entrenchment: Evidence from seasoned
equity offerings
2.1 Introduction ....................................................................... 31
2.2 Sample construction and variable deﬁnition .................................. 35
2.2.1 Sample construction ..................................................... 35
2.2.2' Variable deﬁnition ....................................................... 38
2.3 Empirical tests ..................................................................... 42
2.3.1 Announcement returns .................................................. 42
2.3.2 Multivariate analysis of announcement returns ...................... 43
2.3.3 Multivariate analysis of change in operating performance. . . . . ....45
2.4 Conclusion ......................................................................... 46

Chapter 2 Append1x48

3 Product market competition and ﬁrms’ credit ratings

3.1
3.2
3.3

3.4
3.5
3.6

Introduction ....................................................................... 55
Sample construction ............................................................. 58
Variable deﬁnitions .............................................................. 59
3.3.1 Measuring product market competition .............................. 59
3.3.2 Variables for Tobin’s Q regression .................................... 60
3.3.3 Variables for ﬁrms' credit rating regression .......................... 62
Descriptive statistics .............................................................. 67

The relation between product market competition and Tobin‘s Q. . . . . ....68
The relation between product market competition and ﬁrms’

credit ratings ....................................................................... 70
3.6.1 Product market competition and credit ratings. 71

vi

3.6.2 Product market competition and investment-grade credit rating..73
3.7 The link between product market competition and risky investments. . ...75

3.8 Conclusion .......................................................................... 76
Chapter 3 Appendix ......................................................................... 78

A31 Variable Deﬁnitions ............................................................. 79
Bibliography ................................................................................. 88

vii

1.1

1.2

1.3

1.4

1.5

1.6

1.7

1.8

2.1

2.2

2.3

2.4

2.5

2.6

3.1

3.2

3.3

3.4

3.5

LIST OF TABLES

Sample distribution by issue year ................................................. 22
Industry distribution of sample .................................................... 23
Summary statistics .................................................................. 24

Announcement return comparison of single-class and dual-class SEOS. ...25

SEO announcement abnormal returns and divergence between voting and

cash ﬂow rights ...................................................................... 26
Regression analysis of announcement returns on divergence between voting
and cash ﬂow rights ................................................................. 28
Regression analysis of long-run abnormal buy and hold returns on
divergence between voting and cash ﬂow rights ............................... 29
Factor regressions on portfolios of ﬁrms with SEOS ............................ 30
SEO sample distribution by issue year ........................................... 49
Industry distribution ................................................................ 50
Summary statistics .................................................................. 51
Announcement abnormal returns and board classiﬁcations ................... 52

Regression analysis of announcement returns on board classiﬁcation... . ..53

Regression analysis of change in operating performance on board

classiﬁcation ......................................................................... 54
Credit rating categories ............................................................. 81
Summary statistics .................................................................. 82

Regression analysis of the effect of market competition on Tobin's Q. . ....83
Regression analysis of the effect of market competition on credit ratings..84

The effect of market competition on receiving an investment grade credit
rating ................................................................................... 85

viii

3.6 Probability changes in receiving an investment grade credit ratings. . . . . ....86

3.7 Regression analysis of the effect of market competition on expenditures in
R&D and Advertising .............................................................. 87

ix

Chapter]
Agency problems at dual-class ﬁrms: Evidence from seasoned

equity offerings

1.1 Introduction

Jensen and Meckling (1976) argue that high managerial equity ownership in ﬁrms
foster a better alignment of managerial interests with those of shareholders and
helps in alleviating the agency problem associated with the separation of ownership
and control. They argue that managers bear a larger proportion of the costs of
shirking, consumption of perquisite and other value-destructive actions as their
ownership increases. Their paper focuses primarily on cash flow rights inherent in
equity ownership. However, higher control or voting rights due to increased
ownership can entrench management, resulting in a lower alignment of interests
between managers and minority shareholders and thus increase agency cost
(Gompers, Ishii and Metrick (2006)). Since the effects of cash ﬂow rights and
control rights are not easily separable, researchers face a challenging task in
investigating empirically the effect Of managerial ownership on ﬁrm value

The unique ownership structure of dual-class ﬁrms, which usually implies
signiﬁcant divergence between voting and shareholder rights, allows us to examine
the different effects of cash flow rights and voting (or control) rights. In a typical
dual-class ﬁrm, there are two classes of stocks, which are a publicly traded inferior

class Of stock with one vote per share and a non-publicly traded superior class of

stock with multiple votes per share. The superior class of stock is usually held by
management. A divergence between managerial voting and cash ﬂow rights implies
that managers’ interests are less aligned with shareholders interests and thus
creating incentives for managers to extract private beneﬁts at the expense of
shareholders, ultimately leading to a reduction in ﬁrm value. Consistent with this
line of reasoning, Gompers, Ishii and Metrick (2006) and Masulis, Wang and Xie
(2007) Show that the divergence between voting rights and cash ﬂow rights
exacerbates the agency problems between managers and shareholders.

We explore the effect of misalignment between managers and shareholders
interest on the market reaction to seasoned equity offering (SEO) announcements
and long-run stock performance following 81303. We focus on the 81303 by dual-
class ﬁrms since they provide a unique framework where we can disentangle the
effect of managers’ voting and cash ﬂow rights on ﬁrm value. To the extent that
investors are concerned about agency problems, we expect that market to react
more negatively to SEO announcements as the divergence between managerial
voting and cash ﬂow rights becomes greater. Moreover, the misaligmnent of
interests between managers and shareholders can create managerial incentives to
pursue private beneﬁts and undertake value-destroying investments, which
decreases ﬁrm value. Therefore, we expect the long-run stock performance would
be negatively related to the divergence between voting and cash ﬂow rights.

Previous studies have documented a signiﬁcantly negative stock price
reaction to announcement of SEOs. The existing literature offers several

explanations for this negative reaction. The Leland and Pyle (1977) signaling

theory suggests that sales of shares by insiders signal that they believe that the
shares are overpriced. Myers and Majluf (1984) extend this theory by arguing that
issuing equity should be the least preferred choice of ﬁrms to raise capital since the
mere act of issuing equity conveys a negative signal about the true value of the ﬁrm.
Jung, Kim and Schulz (1996) provide a theory of agency costs for the negative
reaction. They argue that when interests of managers and shareholders are not
aligned, managers may pursue value-destructive investments to increase their
private beneﬁts, which decrease ﬁrm value. Therefore, investors react negatively to
SEO announcements due to the potential concern about misuse of funds raised in
the SEOS.

When it comes to a comparison between single and dual-class ﬁrms,
Megginson, Smart and Zutter (2008) argue that signaling and agency theory
provide conflicting predictions regarding SEO announcement returns. They argue
that signaling theory predicts that markets react more negatively to single-class
SEO announcement since SEOS by single-class ﬁrms make the insiders incur the
cost of diluting voting- rights at a faster rate and thus signal more overvaluation than
SEOS by dual-class ﬁrms. On the contrary, agency theory predicts that market react
more negatively to dual-class SEO announcement because Of the misaligned
interests between managers and shareholders in dual-class ﬁrms. Since both of
these confounding effects could occur simultaneously, the comparison of SEO
market reaction between single and dual class ﬁrms remains ultimately an
empirical question. Megginson, Smart and Zutter (2008) ﬁnd no signiﬁcant

difference in SEO announcement retums between single and dual-class ﬁrms.

We ﬁnd that SEO announcement returns are negatively related to the
divergence between control and cash ﬂow rights. We also ﬁnd that three year
abnormal buy-and-hold returns following SEOS are also negatively related to the
divergence between control and cash ﬂow rights. Finally, we ﬁnd that investment
strategy that bought ﬁrms that have below median value of divergence between
voting rights and cash flow rights and sold ﬁrms that have above median value of
the voting rights and cash ﬂow rights would have earned positive abnormal returns.

Taken as a whole, our results support the view that misalignment of
interests between manager and shareholder caused by divergence between voting
rights and Shareholder rights can partly explain the negative market reactions to
SEO announcements and poor long-run performance following SEOs in dual—class
ﬁrms.

The remainder of this paper is organized as follows: Section 2 describes the
sample construction. Section 3 shows the analysis of SEO announcement returns.
Section 4 shows the analysis of long-run stock performance following 811305.

Finally, we present our conclusions in Section 5.

1.2 Sample construction

We start with the sample of dual-class ﬁrms that Gompers, Ishii and Metrick (2006)
construct'. They construct the dataset from the universe of publicly-traded ﬁrms in
the US. for the period from 1994 to 2002. About 6% of publicly-traded ﬁrms in the

US. in the period have a dual-class structure. In a typical dual-class ﬁrm, there are

 

' We are grateful to Paul Oompers. Joy Ishii, and Andrew Metrick for generously making the data
available.

two classes of stocks — a publicly traded inferior class of stock with one vote per
share and non-publicly traded superior class of stock with multiple votes per share.

We then compile all SEOS for the period of 1994 to 2003 from the
Securities Data Company’s (SDC) new issues database. We require that SEOs must
be for common stock by US. issuers. We also require that the stocks are listed on
NYSE, NASDAQ or AMEX. We use publicly traded inferior class of stocks for our
analysis. Since pure secondary offerings do not produce proceeds to the issuing
ﬁrms, we require that SEOs must include some primary shares in the offering. We
exclude utilities (SIC codes between 4910-4949) and ﬁnancial ﬁrms (SIC codes
between 6000-6999). The resulting SEO sample is then merged with the sample of
dual-class ﬁrms of Gompers, Ishii and Metrick (2006). We further restrict the
resultant sample by allowing only one SEO per ﬁrm per year since statistical
inferences could be inﬂuenced by ﬁrms with multiple SEOs in the same year (Kim
and Pumanandam (2006)). Our ﬁnal sample consists of 145 SEOS by dual-class
ﬁrms.

Table 1.1 reports the distribution of our sample of SEOS by issue year,
along with mean (median) statistics for each issue year and for the full sample. Our
sample ﬁrm has mean (median) market capitalization of $1,377 million ($433
million) at the end of the ﬁscal year prior to the SEO and mean (median) proceeds
raised in the SEO is $186 million ($94 million). The mean (median) offer size,
deﬁned as proceeds divided by market capitalization at the end of the ﬁscal year
prior to the SEO is 27% (19%). This large offer size indicates that the SEOS are

important events for our sample ﬁrms. This table also reports cumulative abnormal

return over a (-1, +1) event window around the SEO announcement date.
Consistent with previous literature, SEO announcement returns are on average
signiﬁcantly negative. Our sample of dual-class ﬁrms, on average, earns a negative
2.5% return around the SEO announcement date.

Table 1.2 reports the industry distribution of the SEOS by dual-class ﬁrms.
We use the 12 industry classiﬁcation of F ama and French obtained from Kenneth
French’s web page to classify the ﬁrms. The frequency of 81303 is not evenly
spread across the industry. Firms belonging to the Telecom industry represent
roughly 24% of our sample whereas Shops, Manufacturing, Healthcare and
Business equipment represent roughly 19%, 12%, 9% and 8% respectively of our
sample. The fact that the Telecom industry represents the highest proportion of
ﬁrms in our sample is not surprising because of their possibly high non-pecuniary
private beneﬁts of control thus inducing founders to establish a dual-class structure
(Gompers, Ishii and Metrick (2006)). Since we include industry ﬁxed effects in the
multivariate regressions of announcement retums, we are not concerned about

time-invariant industry effects.

1.3 Analysis of SEO announcement returns
In this section we explore the inﬂuence of the divergence between voting rights and
cash flow rights on the abnormal announcement returns for SEOS by dual-class

ﬁrms.

1.3.1 Divergence between voting rights and cash flow rights

The key variables in this study are measures of divergence between managers’
voting rights and cash ﬂow rights for dual—class ﬁrms. Following Gompers, Ishii
and Metrick (2006) and Masulis, Wang and Xie (2007), voting rights (VT) are
calculated as aggregate voting rights held by managers as a fraction of all
outstanding voting rights, while cash flow rights (CF) are calculated as aggregate
cash ﬂow rights (or dividend rights) held by managers as a fraction of all
outstanding cash ﬂow rights (or dividend rights). We then deﬁne two measures of
the divergence between voting rights and cash flow rights. The ﬁrst measure is
deﬁned as the wedge between voting rights (VT) and cash ﬂow rights (CF), while
the second measure is deﬁned as the ratio of voting rights (V7) to cash ﬂow rights
(C F). Both measures are designed to increase with divergence between voting
rights and cash ﬂow rights. The greater divergence may give managers more
incentives to pursue private beneﬁts. Table 1.3 shows that managers own mean
(median) voting rights of 64.5% (68.6%), while managers own mean (median) cash
flow rights of 40.8% (39%), which indicates considerable divergence between
voting rights and cash flow rights. Speciﬁcally, the mean (median) Wedge is 23.6%

(25.6%), whereas the mean (median) Ratio is 1.791 (1.548).

1.3.2 Variable deﬁnition

Based on the previous studies (see e.g. Kim and Pumanandam (2006) and Ferreira
and Laux (2007)), we include a number Of variables to control for other
determinants of the SEO announcement returns in multivariate analysis. We control

for various ﬁrm characteristics. We control for ﬁrm size, which is deﬁned as the

logarithm of the total assets. Firm size is used as a proxy for asymmetric
information. Ferreira and Laux (2007) argue that ﬁrm size tends to reduce the
information asymmetry since larger ﬁrms are more like to be under greater scrutiny
by outside investors. We expect ﬁrm size (Log (D1)) to be positively related to
announcement returns. We also control for market-to-book ratio (MB), which is
calculated as the market value of the ﬁrm’s equity plus the book value of assets
minus the book value of equity divided by the book value of assets. Growth ﬁrms
are more likely to have more proﬁtable investment opportunities. Therefore, we
expect market-to-book ratio to be positively related to announcement returns.

We include past returns (Past Return), which is calculated as the ﬁnn’s raw
buy-and-hold return over a period of one year prior to the SEO issue date. Prior
research provides competing arguments on the effect of ﬁrms’ past returns (e.g.
Kim and Pumanandam (2006)). Firms’ past returns can be used as a proxy for the
availability of good projects or a proxy for overvaluation of stocks. Therefore, we
do not have a clear prediction about the relationship between ﬁrms’ past returns
and announcement returns. Myers and Majluf (1984) argue that ﬁnancial slack of a
ﬁrm reduce the problem of adverse selection. To control for the ﬁnancial slack, we
also include a firm’s cash divided by total assets (Cash). Firm leverage is also used
as a control variable. We calculate leverage as the sum of debt in current liabilities
plus long term debt divided by total assets. Leverage can reduce the agency
problems by controlling managerial discretion (Stulz (1990)). However, competing

argument suggests that high leverage can create managerial incentives to take risky

negative-NPV projects (Jensen and Meckling (1976)). Therefore, we do not clearly
predict relationship between leverage and announcement reruns.

We also include some issue characteristics as control variables. Following
Ferreira and Laux (2007), we include Offer Size to control for the economy of scale
effect. Oﬂer Size is deﬁned as the amount of proceeds raised from the SEO divided
by market capitalization at the end of the ﬁscal year prior to the issue. We expect a
positive relationship between the Offer Size and announcement returns. We also
include a dummy variable Secondary that equals one if some secondary shares are
included in the SEO Offering and zero otherwise. In addition to ﬁrm and issue
characteristics, we also control for Accrual since earnings management may affect
investors’ response to the announcement of SEO. Accrual is deﬁned as the
difference between net income and cash flow from operations divided by total
assets (Kim and Pumanandam (2006)).

Table 1.3 reports the summary statistics for the variables employed in the
study. Table 1.3 also includes some control variables used in the analysis of long-
run performance following SEOs, which will be discussed in a later section.
Included are the number of observations, mean, median, standard deviation, 25th
percentile and 75th percentile values for the variables in the analysis. The mean
(median) value of Log (TA) is 6.214 (6.196). Market-to-book ratio (MB) has a mean
of 2.156 and a median of 1.624, with 25th and 75th percentile values of 1.236 and
2.448, respectively. The mean market capitalization is 1400, while the median
market capitalization is 384, which indicates the distribution is right-skewed

(positive skewness). The amount of proceeds as a percentage of pre-issue market

capitalization (Offer Size) has a mean and median of 27.5% and 19.1%,
respectively. The average ﬁrm in our sample has Leverage of 38.8% and Cash of
10.5%. On average, 48.3% of the SEOs include some secondary shares in the

Offerings.

1.3.3 Univariate Analysis

To determine the announcement date of the SEO, we search all publications
included in Factiva including Wall Street Journal and the Dow Jones News
Retrieval Services during the two months before the ﬁling date. If we ﬁnd any
news publication of a ﬁrm’s plan of an SEO, we consider the publication date as
the announcement date of the SEO. Otherwise, the ﬁling date from the SDC
database is used as the announcement date. We calculate cumulative abnormal
returns for all our empirical analyses using a standard market model with
parameters estimated over days -249 and -50 relative to the announcement date.

We start our empirical analysis with a brief comparison of SEO
announcement returns between the sample of single-class and dual-class ﬁrms and
the results are reported in Table 1.4. Panel A reports the results for the (-1, +1)
event window and Panel B reports the results for the (-l, 0) event window. We
apply the same restrictions in selecting the sample Of single class ﬁrms that we
required for the dual-class ﬁrms. We ﬁrst extract all SEOS from the SDC database
for the period 1994-2003. We require that SEOS be for common stock by US.
issuers, listed on NYSE, NASDAQ or AMEX. We exclude pure secondary

offerings and remove all ﬁnancial and utility ﬁmis from our sample. Finally, we

10

restrict our sample to only one SEO per ﬁrm per year. Since single-class ﬁrms are
not the primary focus of this study, we use the ﬁling dates from the SDC database
as the announcement dates considering the large sample size of the SEOs by single-
class ﬁrms. However, previous research has evinced that announcement date
coincide with the ﬁling date from the SDC database for a large percentage of SEOS
(Ferreira and Laux (2007)).

Results in Table 1.4 shows that SEO announcement returns are negative and
statistically signiﬁcant for all SEOs as well as for single-class and dual-class ﬁrms.
However, the mean and median announcement return differences between single
and dual class ﬁrms are not statistically signiﬁcant. The result is consistent with
Megginson, Smart and Zutter (2008) who argue that signaling theory and agent
theory operate in opposite way, which leads to ambiguity in predicting the
difference in SEO announcement returns between single-class and dual-class ﬁrms.
The rest Of the analysis in this paper focuses solely on the sample of dual-class
ﬁrms described previously.

Table 1.5 reports cumulative abnormal returns (CARS) around SEO
announcement dates for our sample of dual-class ﬁrms. Panel A and Panel B report
the result for the (-1, +1) event window and (-1, 0) event window, respectively. The
SEO announcement returns for our sample ﬁrms are negative and statistically
signiﬁcant in both event windows. We further divide the dual-class sample into
sub-samples based on divergence of voting rights and cash flow rights. A ﬁrm is
categorized in the high wedge (ratio) group if the wedge (ratio) is at or above the

median, otherwise, the ﬁrm is in the low wedge (ratio) group. Panel A shows that

11

ﬁrms with low Ratio, on average, outperform ﬁrms with high Ratio by 1.7%.,
which is statistically signiﬁcant at 10% level. However, the mean and median
differences in returns between ﬁrms with low Wedge and high Wedge are not
statistically signiﬁcant.

Panel B reports that ﬁrms with a high Wedge experience a mean (median)
CAR of -1.8% (-1.6%), which is signiﬁcantly greater in magnitude than the mean
(median) of -0.5% (less than -0.0001%) for ﬁrms with low Wedge. Also, we ﬁnd
that ﬁrms with high Ratio get a signiﬁcantly more adverse median stock price
response (-1.6%) than that for ﬁrms with low Ratio (less than -0.0001%). However,
we ﬁnd no statistically signiﬁcant difference between mean announcement return
of ﬁrms with high Ratio and low Ratio. In summary, the results in Table 1.5
suggest that SEO announcement returns for dual-class ﬁrms are negative. Moreover,
there is some weak evidence that the greater the divergence between managers’
voting rights and cash flow rights, the more negative the SEO announcement
returns for dual-class ﬁrms. This evidence indicates that investors are concerned
about the agency problems caused by divergence between voting rights and cash
ﬂow rights when dual—class ﬁrms issue equity. We need additional multivariate

analysis before we reach our conclusions.

1.3.4 Multivariate analysis
In this section, we investigate the effect of divergence between voting rights and
cash ﬂow rights on the SEO announcement returns in a multivariate setting. Table

1.6 reports the results of regression analysis of SEO announcement returns on the

12

measure of divergence between voting rights and cash ﬂow rights after controlling
for other explanatory variables. We include explanatory variables following Kim
and Pumanandam (2006) and Ferreira and Laux (2007). We also include year and
industry dummies (based on 2-digit SIC). The dependent variable is the cumulative
abnormal returns (CARS) during (-1, +1) event window surrounding the SEO
announcements. The estimated coefﬁcient on Leverage is negative and statistically
signiﬁcant, which supports the view that high leverage can create managerial
incentives to take risky negative-NPV projects (Jensen and Meckling (1976)). The
coefﬁcient on Secondary is also negative and statistically signiﬁcant. The result
indicates that the market react adversely to the sale of securities by s and large
shareholders, which is consistent with Kim and Pumanandam (2006).

The key explanatory variables are Wedge and Ratio. A higher value of
Wedge or Ratio indicates greater divergence between voting rights and cash ﬂow
rights and thus more misalignment of manager and shareholder interests. Therefore,
we expect to see a negative coefﬁcient on these variables to the extent that
investors are concerned about agency problems. Consistent with our expectation,
Model 1 shows that the coefﬁcient estimate of Wedge is negative and statistically
signiﬁcant. The result suggests that misalignment of manager and shareholder
interests is associated with lower announcement returns. Model 2 includes squared
term of Wedge and repeats the analysis to examine whether there is a nonlinear
effect of the divergence between voting rights and cash ﬂow rights on

announcement returns. The estimated coefﬁcient on Wedge is continuously

13

negative and statistically signiﬁcant but the estimated coefﬁcient on Wedge: is not
statistically signiﬁcant.

In Model 3, we replace the variable Wedge with the variable Ratio and
Model 4 includes the squared term of Ratio. In Model 3, the estimated coefﬁcient
on Ratio is insigniﬁcant. However, when we include Ratio: in Model 4 and repeat
the analysis, we ﬁnd that the coefﬁcient estimates of Ratio and Ratio: are -0.039
and 0.007, respectively, and are statistically signiﬁcant at 10% level. These
estimated coefﬁcients indicate that announcement returns decrease until Ratio
reaches about 2.79 and then announcement returns start to increase with further
increases in Ratio. Overall, our results in Table 1.6 suggest that divergence between
voting rights and cash ﬂow rights are negatively related to SEO announcement

returns .

1.4 Analysis of long-run stock performance following SEOs

1.4.1 Long-run abnormal returns

This section examines the long-run abnormal retums following equity issues. We
use a matching ﬁrm benchmark to compute abnormal returns. We choose matching
ﬁrms using a procedure analogous to that used by Loughran and Ritter (1995) and
Lee (1997). Non-issuing matching ﬁrms are chosen on the basis of size, book-to-
market ratios and prior annual return for ﬁrms with available book-to-market ratios.
Otherwise, non-issuing matching ﬁrms are chosen on the basis of size and prior
annual return. To be included in the pool of candidate matching ﬁrms, the ﬁrms

should be listed in CRSP. TO be consistent with our sample selection criteria, we

14

exclude utilities and ﬁnancial ﬁrms and we just include the ﬁrms for US. common
stocks listed on NYSE, NASDAQ, or AMEX.

The following describes the non-issuing matching ﬁrm selection procedure on the
basis of size, book-to-market ratios and prior annual return for ﬁrms with available
book-to-market ratios. At the end of each month, we ﬁrst sort the pool of candidate
matching ﬁrms on the NYSE into quintiles based on their market capitalization.
The market capitalization is calculated as the product of shares outstanding and the
closing price of the stock at the end of each month. We then place the pool of
candidate matching ﬁrms on the NASDAQ and AMEX ﬁrms into the appropriate
size quintiles that we deﬁned using NYSE stocks. For issuing ﬁrms, the market
capitalization is calculated at the offering date. Within each size quintile, we sort
ﬁrms into quintiles based on their book-to-market ratios. The book-to-market ratio
is calculated as the ratio of the book value of equity to the market value of equity at
the end of each month. This gives us a total of 25 portfolios sorted on the basis of
size and book-to-market ratios. The reported book value of equity for a given ﬁscal
year is not used until at least four months have elapsed following the end of the
ﬁscal year. For example, ﬁrms that have ﬁscal years ending on November 30, the
new book value of equity is used on or after March 31. Finally, we further divide
the resulting 25 portfolios into quintiles based on ﬁrms’ prior return resulting in a
total of 125 (5 x 5 x 5) portfolios at the end of each month. We calculate prior
return by cumulating monthly returns over the previous one year. From the

portfolio of the same size, book-to-market ratio, and prior annual return at the time

15

of issue, we then choose the ﬁrm whose prior return is closest to that of the issuing
ﬁrm as the matching ﬁrm.

If the book—to-market ratio is not available for an issuing ﬁrm, we choose
our non-issuing matching ﬁrm on the basis of size and prior return in the following
way. First we sort the pool of candidate matching ﬁrms on the NYSE into deciles
based on their market value of equity at the end of each month. We then place the
pool of candidate matching ﬁrms on the NASDAQ and AMEX ﬁrms into the
appropriate size deciles that we deﬁned using NYSE stocks. Within each size
decile, we sort ﬁrms into quintiles based on prior annual return. This gives us a
total of 50 (10 x 5) portfolios at the end of each month. From the portfolio of the
same size and prior return at the time of issue, we then choose the ﬁrm whose prior
return is closest to that of the issuing ﬁrm as the matching ﬁrm.

We use three-year buy-and-hold returns to measure long-run performance

of issuing ﬁrms and non-issuing matching ﬁrms. Let r" denote the daily return to

stock i on date I and T denotes the last date of holding period, which is either the
end date of three year anniversary or the delist date, whichever comes ﬁrst. The

buy-and-hold return R, is then deﬁned as:

Rit = ﬁ<1+rit)—1

t=1

We calculate three-year buy-and-hold return for each issuing ﬁrm. If an issuing
ﬁrm is delisted before the three year anniversary of its issue date, the buy-and-hold
return truncate on the issuing ﬁrm’s delisting date. For each matching ﬁrm, we use

the same holding period as the issuing ﬁrm to calculate buy-and-hold retums. 1f a

16

matching ﬁrm is delisted before the holding period of the issuing ﬁrm, we splice
the next best matching ﬁrm as of the original matching date into the calculation of
the buy-and-hold return from the day after delisting date. If a matching ﬁrm issues
equity before the end of holding period of issuing ﬁrm, the matching ﬁrm is treated
as being delisted on its issue date and the next best matching ﬁrm is spliced into the
calculation of the buy-and-hold return from the day after delisting date (issue date).
Whenever the spliced ﬁrm is delisted again, we repeat the same procedure above.
The three-year abnormal return following an SEO is then calculated as the
difference between buy-and-hold return of issuing ﬁrm and non-issuing matching
ﬁrm.

Table 1.7 reports the results of the regression analysis of the three—year
abnormal returns on divergence between voting rights and cash flow rights. The
control variables are market-to-book ratio (MB), market capitalization (Market
Capital), one year return prior to issue (Past Return), Oﬂ'er Size, and market
adjusted capital expenditures (CAPEX). Model 1 reports the estimates using Wedge
to measure divergence between voting rights and cash ﬂow rights. The coefﬁcient
on Wedge is negative and statistically signiﬁcant. The result indicates that
misalignment of manager and shareholder interests results in Signiﬁcantly worse
ﬁ'nrr performance following SEOS. In Model 2, we replace the variable Wedge with
the variable Ratio and repeat the analysis. The result in Model 2 shows that the
estimated coefﬁcient on Ratio is negative and statistically signiﬁcant. In summary,
the results in Table 1.7 suggest that divergence between voting rights and cash ﬂow

rights lower the long-run performance following SEOs.

l7

1.4.2 Factor regressions
In each of calendar month of our sample, a portfolio is formed that includes those
firms that had an SEO in the previous 36 months. The calendar time returns on

these portfolios are then used to estimate the following four-factor model by

Carhart (1997):
R, = a + ATMKT, + ﬂz’kSMB, + ,B3*HML, + ,84*UMD,
where R, is the excess return in month t, MKT, is the excess return on the market

portfolio in month t, SMB,, HML,, and UMD, are returns on the portfolios designed
to capture size, book-to-market, and momentum in month t. We report the result
based on value-weighted portfolios. Model 1 shows the result of estimating above
model where the dependent variable R, is the monthly return difference between
portfolio of ﬁrms that have above median value of Wedge and ﬁrms that have
below median Wedge. Therefore, the alpha here is the abnormal return on a zero-
investment strategy that is long in ﬁrms that have above median value of Wedge
and short in ﬁrms that have below median value of the Wedge. The estimate of the
intercept, or,, is negative and statistically signiﬁcant at 5% level. The result indicates
that return difference between above-median Wedge and below-median Wedge
samples, which is measured by the return of the zero-investment portfolio, is
negative and statistically different from zero.

In Model 2, we replace the variable Wedge with the variable Ratio and
repeat the analysis. In Model 2, the dependent variable R, is now the monthly return
difference between portfolio of ﬁrms that have above median value of Ratio and

firms that have below median Ratio. Therefore, the alpha is now the abnormal

l8

return on a zero-investment strategy that is long in ﬁrms that have above median
value of Ratio and short in ﬁrms that have below median value of the Ratio. We
ﬁnd that the alpha is again negative and statistically signiﬁcant at 5% level. This
ﬁnding suggests that return difference between above-median Ratio and below-
median Ratio samples is negative and statistically signiﬁcant. Overall, the results of
Table 1.8 indicate a negative relation between the misalignment of interests

between managers and shareholders and the long-run stock performance following

SEOs.

1.5 Conclusion
The unique ownership structure of dual-class ﬁrms, which is the divergence
between managers’ voting and shareholder rights, allows us to investigate agency
problems in SE05. Agency theory suggests that misalignment of interests between
managers and shareholders can create managerial incentives to conduct value-
destroying investments to extract private beneﬁts at the expense of shareholders,
which decrease ﬁrm value. To the extent that investors are concerned about agency
problems, the greater the divergence between voting rights and shareholder rights
the more negatively the market react to the SEO announcements by dual-class
ﬁrms.

We ﬁnd that market responds more adversely to SEO announcement as the
divergence between voting rights and cash flow rights increases. We also ﬁnd that
ﬁrms experience worse three year abnormal returns following SEOS as the

divergence between voting rights and cash ﬂow rights increases. Finally, we ﬁnd

19

that investment strategy that bought ﬁrms that have below median value of
divergence between voting rights and cash ﬂow rights and sold ﬁrms that have
above median value of divergence between the voting rights and cash flow rights
would have earned positive abnormal returns during the sample period.

In summary, our results support the view of agency problems in SEOS.
Speciﬁcally, misalignment of interests between managers and shareholders caused
by divergence between voting rights and shareholder rights can partly explain the
negative market reaction to SEO announcements and poor long-run performance

following SEOS in dual-class ﬁrms.

20

Chapter 1 Appendix

21

Table 1.1
Sample distribution by issue year

This table reports the distribution of our sample of seasoned equity issues (SEOs) by issue year, along
with mean (median) statistics for each issue year and for the full sample. Proceeds represent the total
amount of money raised through the SEOs in millions of dollars. Offer size represents the proceeds raised
in the SEO as a fraction of its market capitalization at the end of the ﬁscal year prior to issue. Pre-mktcap
denotes the market capitalization of the issuing ﬁrm, in millions of dollars, at the end of the ﬁscal year
prior to the security issuance. CAR {-1, +1 ) measures cumulative abnormal return over the event window
(-1 , +1) days around the announcement date.

 

 

 

Number Percentage of .
Year of Proceeds Offer Size Pre-mktcap CAR (-1 ,+ 1)
SEOs sample

1994 2 1.4 144 0.16 933 -0.004
(144) (0.16) (933) (-0.004)
1995 6 4.1 94 0.32 489 ~0.007
(54) (0.22) (249) (-0.046)
1996 22 15.2 77 0.45 286 -0.017
(74) (0.34) (183) (-0.025)
1997 19 13.1 92 0.28 436 -0.024
(80) (0.23) (312) (-0.009)
1998 20 13.8 119 0.27 662 -0.019
(63) (0.23) (245) (-0.021)
1999 19 13.1 282 0.19 2329 -0.012
(174) (0.14) (953) (-0.014)
2000 16 1 1.0 315 0.22 2054 -0.044
(224) (0.15) (868) (-0.048)
2001 21 14.5 283 0.19 2626 -0.040
(129) (0.15) (910) (-0031)
2002 11 7.6 239 0.25 2531 -0.018
(154) (0.12) (896) (-0.017)
2003 9 6.2 141 0.34 671 -0.037
(90) (0.24) (426) (-0033)
Total 145 100.0 186 0.27 1377 -0.025
(94) (0.19) (433) (-0.021)

 

22

Table 1.2
Industry distribution of sample

This table reports the industry distribution of seasoned equity issues (SEOS)
in our sample. The SEO ﬁrms are classiﬁed into industries based on the 12
industry classiﬁcation by Fama and French.

 

 

 

Industry Number of SEOS
Business equipment 12
Chemicals 1
Consumer durables 3
Consumer nondurables l 1
Healthcare 14
Manufacturing 1 8
Shops 28
Telecom 35
Other 23
Total 145

 

23

Table 1.3
Summary statistics

This table reports summary statistics for variables employed in the study. VT is deﬁned as the
aggregate voting rights held by managers and directors as a fraction of all outstanding voting rights.
CF is deﬁned as the aggregate cash ﬂow rights held by managers and directors as a fraction of all
outstanding cash ﬂow rights. Wedge is calculated as the difference between VT and CF. Ratio is
calculated as the ratio of VT to CF. Log (TA) denotes the log of total assets (data 6) in the ﬁscal year
prior to the issue. MB is deﬁned by the market value of assets divided by book value of assets,
which is calculated as (data 6 + data 199*data 25 - data 60)/data 6. Market Capital is the product of
number of shares outstanding on the offering date and the ﬁrm’s offering date closing price,
measured in millions of dollars. Offer Size denotes the ratio of total proceeds from the SEO to the
ﬁrm’s market capitalization at the end of the ﬁscal year prior to the issue. Past Return is the raw buy—
and-hold return over a period of one year prior to the SEO issue date. Leverage is the ratio of the
sum of debt in current liabilities (data34) plus long term debt (data 9) to total assets (data6) in the
ﬁscal year prior to the issue. Cash denotes cash and marketable securities (data 1) as a fraction of
total assets (data 6) in the ﬁscal year prior to the issue. Accrual is the ratio of the difference between
cash ﬂow from operating activities and net income to total assets in the ﬁscal year prior to the issue,
which is calculated as (data 172 — data 308) / data 6. Secondary is a dummy variable that takes the
value of one if some secondary shares are included in the SEO offering and zero otherwise. C APEX
is the difference in capital investments as a fraction of total assets (data 128/data 6) for the issuing
ﬁrm and the median ﬁrm in the same industry (based on two-digit SIC code) as the issuing ﬁrm for
the ﬁscal year following the issue.

 

 

N Mean Std. Ql Median Q3
VT 145 0.645 0.273 0.491 0.686 0.861
CF 145 0.408 0.223 0.211 0.390 0.560
Wedge 145 0.236 0.187 0.075 0.256 0.360
Ratio 145 1.791 0.901 1.116 1.548 2.141
Log (TA) 145 6.214 1.437 5.250 6.196 7.165
MB 145 2.156 1.563 1.236 1.624 2.448
Market Capital 145 1400 31 11 197 384 l 174
Offer Size 144 0.275 0.243 0.130 0.191 0335
Past Return 145 0.956 1.395 0.183 0.578 1.043
Leverage 143 0.388 0.280 0.171 0.352 0.557
Cash 145 0.105 0.169 0.012 0.030 0.125
Accrual 145 ~0.048 0.094 -0.077 -0.040 -0.001
Secondar
(dummy)y ’45 0'483 0'5 01 0.000 0.000 1.000
CAPEX 132 0.025 0.107 -0.016 -0.001 0.030

 

24

Table 1.4
Announcement return comparison of single-class and dual-class SEOs

This table reports cumulative abnormal returns (CARS) around SEO announcement dates.
Abnormal returns are computed using a standard market model with parameters estimated
over days -249 and -50 relative to the announcement date. Panel A shows CARS for the
event window (-1, +1) days and Panel B shows CARS for the event window (-1, 0) days. P-

values in parentheses are based on t-tests for difference in means and Wilcoxon rank-sum
test for difference in medians.

Panel A : Event window: (-1, +1)

 

 

Mean Median N
All $13.05 -0.024 -0.025 2194
Dual-class -0.025 -0.021 144
Single-class -0.024 -0.026 2050
(A) - (B) -0.001 0.004
(0.851) (0.999)

 

Panel B : Event window: (-1, 0)

 

 

Mean Median N
All SEOS -0.013 -0.014 2194
Dual-class -0.011 -0.008 144
Single-class -0.013 -0.014 2050
(A) - (B) 0.002 0.006
(0.652) (0.371)

 

25

Table 1.5
SEO announcement abnormal returns and divergence between voting and cash
ﬂow rights

This table reports cumulative abnormal returns (CARS) around SEO announcement dates. Wedge is
the difference between VT (voting rights) and CF (cash ﬂow rights). Ratio is the ratio of VT (voting
rights) to CF (cash flow rights). A ﬁrm is categorized in the high wedge (ratio) group if the wedge
(ratio) is at or above the median, otherwise, the ﬁrm is in the low wedge (ratio) group Abnormal
returns are computed using a standard market model with parameters estimated over days -249 and -50
relative to the announcement date. Panel A shows cumulative abnormal returns (CARS) for the event
window (-1, +1) days and Panel B shows CARS for the event window (-1, 0) days. P-values in
parenthesis for all SEOs are based on t-tests for the mean and signed rank test for the median. P-values
in parenthesis for tests of differences are based on t-tests for the mean and a Wilcoxon rank-sum test
for the median.

Panel A : Event window: (-1. +1)

 

Mean Median N

All SEOs -0.025 -0.021 144
(0.000) (0.000)

Wedge;

High Wedge (A) -0.030 -0.025 73
(0.000) (0.000)

Low Wedge (B) ~0.019 -0.016 71
(0.014) (0.004)

(A) - (B) -0.012 -0.010
(0.223) (0.316)

Ratio:

High Ratio (A) -0.033 -0.025 73
(0.000) (0.000)

Low Ratio (B) -0.016 -0.012 71
(0.040) (0.018)

(A) - (B) -0.017 -0.013
(0.090) (0.078)

 

26

Table 1.5

Panel B : Event window: (-1, 0)

(Continued)

 

 

Mean Median N

All SEOS -0.01 1 -0.008 144
(0.005) (0.002)

Wedge;

High Wedge (A) -0.018 -0.016 73
(0.001) (0.000)

Low Wedge (B) -0.005 0.000 71
(0.453) (0.529)

(A) - (B) -0.013 -0.016
(0.097) (0.035)

Ratio;

High Ratio (A) -0.017 -0.016 73
(0.001) (0.000)

Low Ratio (8) -0.005 0.000 71
(0.397) (0.614)

(A) - (B) -0.012 -0.016
(0.144) (0.025)

 

27

Table 1.6
Regression analysis of announcement returns on divergence
between voting and cash ﬂow rights

This table reports the results of the regression analysis of cumulative abnormal returns (CARS) on
divergence between CF (cash ﬂow rights) and VT (voting rights). The dependent variable is the
cumulative abnormal returns for SEO ﬁrms during the event window (-1, +1) days surrounding the
announcement date. Wedge is the difference between VT (voting rights) and CF (cash ﬂow rights).
Ratio is the ratio of VT (voting rights) to CF (cash ﬂow rights). Wedge: and Ratio2 are the squared
terms of Wedge and Ratio, respectively. Log (TA) denotes the log of total assets (data 6) in the ﬁscal
year prior to the issue. MB is deﬁned by the market value of assets divided by book value of assets,
which is calculated as (data 6 + data 199*data 25 - data 60)/data 6 in the ﬁscal year prior to the issue.
Leverage is the ratio of the sum of debt in current liabilities (data 34) plus long term debt (data 9) to
total assets (data 6) in the ﬁscal year prior to the issue. Cash denotes cash and marketable securities
(data 1) as a fraction of total assets (data 6) in the ﬁscal year prior to the issue. Past Return is the raw
buy-and-hold return over a period of one year prior to the SEO issue date. Accrual is the ratio of the
difference between cash ﬂow from operating activities and net income to total assets in the ﬁscal year
prior to the issue, which is calculated as (data 172 — data 308) / data 6. Oﬂer Size denotes the ratio of
total proceeds from the SEO to the ﬁrm's market capitalization at the end of the ﬁscal year prior to
the issue. Secondary is a dummy variable that takes the value of one if the SEO includes some
secondary shares in the offering and zero otherwise. The coefﬁcients on the intercepts are not reported.
All models include industry dummies (based on two-digit SIC code) and year dummies, whose
coefﬁcient estimates are also not reported. P-values in parenthesis are adjusted for heteroscedasticity.

 

 

 

Model 1 Model 2 Model 3 Model 4
Wedge -0.064 -0. 123
(0.064) (0.069)
Wedge: 0.119
(0.234)
Ratio -0.006 -0.039
(0.438) (0.082)
Ratio: 0.007
(0.082)
Log (TA) -0.002 -0.001 -0.001 -0.002
(0.818) (0.832) (0.929) (0.791)
MB -0.008 -0.009 -0.008 -0.009
(0.165) (0.149) (0.172) (0.135)
Leverage -0.05 1 -0.049 -0.056 -0.052
(0.060) (0.066) (0.044) (0.061)
Cash 0.026 0.034 0.037 0.039
(0.582) (0.483) (0.438) (0.419)
Past Return -(.).006 -0.006 -0.005 -0.005
(0.189) (0.210) (0.288) (0.279)
Accrual 0.000 0.003 0.003 0.005
(0.998) (0.968) (0.966) (0.948)
Offer Size 0.015 0.014 0.011 0.012
(0.621) (0.633) (0.738) (0.675)
Secondary 41.026 -0.026 -0.()27 -0.027
(0.073) (0.070) (0.003) (0.063)
N 141 141 141 141
R-Squared 0.421 0.426 0.401 0.416

 

28

Table 1.7
Regression analysis of long-run abnormal buy and hold returns on divergence
between voting and cash ﬂow rights

This table reports the results of the regression analysis of 3-year abnormal returns on divergence
between CF (cash ﬂow rights) and VT (voting rights). Abnormal return is calculated as the
difference between the buy-and-hold return of the sample ﬁrm and the buy-and-hold return of a
matching ﬁrm by size, book-to-market ratio and prior annual return when book-to-market ratio is
available or by size and prior annual return if book-to-market ratio is unavailable. Wedge is the
difference between VT (voting rights) and CF (cash ﬂow rights). Ratio is the ratio of VT (voting
rights) to CF (cash ﬂow rights). MB is deﬁned by the market value of assets divided by book value
of assets, which is calculated as (data 6 + data 199*data 25 - data 60)/data 6 in the ﬁscal year prior
to the issue. Market Capital is the product of number of shares outstanding and closing price on the
offering date, measured in millions of dollars. Past Return is the raw buy-and-hold return
calculated over one year period prior to the issue date. Offer Size denotes the ratio of total proceeds
from the SEC to the ﬁrm’s market capitalization at the end of the ﬁscal year prior to the issue.
C APEX is the difference in capital investments as a fraction of total assets (data 128/data 6) for the
issuing ﬁrm and the median ﬁrm in the same industry (based on two-digit SIC code) as the issuing
ﬁrm for the ﬁscal year following the issue. The coefﬁcients on the intercepts are not reported. P-
values are adjusted for heteroscedasticity.

 

 

 

 

Model 1 Model 2

Estimate p-Value Estimate p-Value
Wedge -l .248 0.073
Ratio -0.323 0.044
MB 0.052 0.647 0.062 0.568
Market Capital -0.000 0.078 -0.000 0.079
Past Return 0.082 0.480 0.102 0.370
Offer Size -0.418 0.271 -0.458 0.234
CAPEX 0.125 0.856 -0.l35 0.847
N 132 132
R-Squared 0.016 0.022

 

29

Table 1.8
Factor regressions on portfolios of ﬁrms with SEOs

This table reports calendar-time factor regressions of portfolios consisting of dual-class ﬁrms
that have conducted SEOs during a period of 36 months prior to the month of portfolio
formation. The four-factor model of Carhart (1997) is estimated based on the following model:

R, = a + awn + BfSMB + B3*HML + ,Bﬁ‘UMD

where R, is the excess return in month t, MK T, is the excess return on the market portfolio in
month t, SMB,, HML,, and UMD, are returns on the portfolios designed to capture size, book-to-
market, and momentum in month t. We report the result based on value-weighted portfolios. In
Model 1, the dependent variable R, is the monthly return difference between portfolio of ﬁrms
that have above median value of Wedge and ﬁrms that have below median Wedge. Wedge is
deﬁned as the difference between VT (voting rights) and CF (cash flow rights). The zero-
investment portfolio is a portfolio that is long in ﬁrms that have above median value of Wedge
and short in ﬁrms that have below median value of Wedge. 1n Model 2, the dependent variable
R, is the monthly return difference between portfolio of ﬁrms that have above median Ratio and
ﬁrms that have below median Ratio. Ratio is deﬁned as the ratio of VT (voting rights) to CF
(cash ﬂow rights). The zero-investment portfolio is a portfolio that is long in ﬁrms that have
above median value of Ratio and short in ﬁrms that have below median value of Ratio. P-
values are adjusted for heteroscedasticity.

 

 

 

Model 1 Model 2

Estimate p-Value Estimate p-Value
a -0.013 0.036 -0.011 0.032
Rm - R, 0.035 0.766 0.178 0.140
SMB 0.057 0.657 0.217 0.091
HML 0.057 0.743 0.335 0.041
UMD -0.021 0.857 -0.020 0.831
R-Sqared 0.002 0.030

 

30

Chapter 2
Classiﬁed boards and managerial entrenchment: Evidence from

seasoned equity offerings

2.1 Introduction

This paper provides evidence that poor governance in the form of classiﬁed boards
can partly explain the negative market reaction to seasoned equity offering (SEO)
announcements and the operating performance decline following SEOS. Previous
studies have documented a signiﬁcant negative stock price reaction to
announcement of SEOS. Myers and Majluf (1984) propose adverse selection
problem. They argue that ﬁrms are more likely to issue equity when the equity is
overvalued in the presence of information asymmetry between managers and
outside investors,. Thus, the announcement of an equity offering conveys negative
information about ﬁrm value. This is the adverse selection problem.

Another explanation for this negative reaction is agency conﬂicts between
managers and Shareholders. This agency explanation, formally introduced by lung,
Kim and Stulz (1996) argues that when managerial interests are misaligned with
shareholder interest, managers may undertake value-destroying investments in
order to increase their private beneﬁts of control. Such misuse of funds raised
through equity offerings, if anticipated by investors, could be a possible

explanation for the negative announcement reaction to seasoned equity offerings.

31

The notion that SEO ﬁrms might sub-optimally invest the proceeds raised
through an offering can be traced back to the free cash ﬂow hypothesis proposed by
Jenson (1986). Jensen (1986) suggests that a managerial tendency to over-invest is
a direct result of empire building and compensation considerations since larger
ﬁrms offer more private beneﬁts and compensation to their executives. The agency
costs of free cash ﬂow are widely documented in the literature. Jensen (1986)
points out the overinvestment problem in the petroleum industry had occurred as a
direct result of excess free cash ﬂow generated due to high oil prices. Fu (2006)
tests the overinvestment hypothesis for SEO ﬁrms and concludes that the free cash
ﬂow problems increase aﬁer a ﬁrm has gone through a seasoned equity offering.

Recent empirical researches on this topic have outlined the importance of
corporate governance in understanding the negative reaction to seasoned equity
offerings. Ferreira and Laux (2007) provide evidence that independent directors
acting as effective monitors not only help in preventing misuse of ﬁmds raised
through an SEO, but also help in reducing adverse selection which is often cited as
one of the standard explanation for negative announcement returns to SEOs. Kim
and Pumanandam (2006) argue that misaligned interest between managers and
shareholders is an important determinant of market’s reaction to SEO and show
that SEO announcement returns are positively related to the sensitivity of
managers’ wealth to stock price movement.

This paper extends the above literature by looking at the role of board
structure (classiﬁed versus unitary) in explaining part of the negative performance

of ﬁrms going through a seasoned equity Offering. A classiﬁed board is a board

32

structure that staggers the annual election of board of directors. In a non-classiﬁed
or unitary board, directors are elected for one year terms at the ﬁrm’s annual
meeting. In contrast, a classiﬁed board is a structure of board of directors in which
every year only a fraction of the directors are elected, each for multiyear terms.
Usually, a classiﬁed board has three classes of directors, which is the largest
permissible number of classes in most states of incorporation. Although classiﬁed
boards have encountered growing resistance from activist shareholders and
institutional investors during the past decade, a majority of American corporations
still utilize such a board structure. Koppes, Ganske and Haag (1999) argue that
classiﬁed elections encourage board independence and increase the effectiveness of
directors in their role as monitors. This also ensures board stability as the majority
of directors serving at any given time would have prior experience as directors thus
providing in-depth knowledge of the functioning of the ﬁrm and the industry as a
whole. Finally, a classiﬁed board discourages short-termism by allowing directors
to focus on long term strategies and enhance the ﬁrm’s ability to create value.

The empirical evidence on classiﬁed boards, however, portrays a dismal
picture of their effectiveness. Babchuk and Cohen (2005) provide evidence that
classiﬁed boards are associated with a lower ﬁrm value. Faleye (2007) shows that
classiﬁed boards reduce director effectiveness, leading to managerial entrenchment
and therefore resulting in destruction of shareholder value. Richardson (2006) ﬁnds
evidence that, ﬁrms utilizing a classiﬁed board structure are associated with higher

levels of overinvestment Of free cash ﬂow.

The above lines of reasoning allow us to construct two alternate testable
hypotheses regarding the market reaction to SEO announcements. The ﬁrst
hypothesis which we will refer to as the monitoring hypothesis posits that if
classiﬁed boards more efﬁciently monitor managers and effective in preventing
them from misuse of funds raised in SEOS, then we would expect a less negative
market reaction to SEO announcements and higher operating performance
following SEOS for ﬁrms with classiﬁed boards relative to ﬁrms with unitary
boards. On the other hand, if classiﬁed boards entrench managers and are
ineffective in preventing them from misuse of funds raised in SEOS, we would
expect a more negative market reaction to SEO announcements and lower
operating performance following SEOS compared to ﬁrms with a unitary board
structure. We will refer to this as the entrenchment hypothesis.

Our paper contributes to the literature by demonstrating that poor
governance in the form of classiﬁed boards can partly explain the negative market
reaction to seasoned equity offering (SEO) announcements. The empirical results in
this paper are consistent with the entrenchment hypothesis discussed above. The
main ﬁndings of our paper are the following: Our analysis of cumulative abnormal
returns (CARS) around SEO announcements shows that ﬁrms having a classiﬁed
board structure have signiﬁcantly lower CARS than ﬁrms having a unitary board
structure. Also, our results from the multivariate regression of CARS on board
classiﬁcation Show that announcement returns of SEOs are more negative for ﬁrms
having a classiﬁed board structure compared to ﬁrms having a unitary board after

controlling for ﬁrm and issue characteristics. Moreover, market reacts negatively to

34

the capital expenditures by ﬁrms having a classiﬁed board structure following the
SEOS. Result from the regression analysis of change in operating performance
suggests that ﬁrms with classiﬁed boards earn 4.7% lower abnormal operating
performance compared to ﬁrms with unitary boards.

The remainder of the paper is organized as follows: Section 2 discusses the
sample construction and deﬁnes the variables used in the multivariate regression.
Section 3 discusses the empirical tests and interprets the results and Section 4

concludes.

2.2 Sample construction and variable deﬁnition

2.2.1 Sample construction

The data for this study is obtained from multiple sources. We obtain data on SEOs
from the Thompson Financial Securities Data Corporation (SDC) database for the
period 1995-2002. We require that SEOS must be for common stock by US.
issuers and the stocks are listed on NYSE, NASDAQ or AMEX. To be included in
our sample, the SEOS must include some primary shares in the offering since pure
secondary offerings do not produce proceeds to the issuing ﬁrm. We remove
utilities (SIC codes between 4910-4949) and ﬁnancial ﬁrms (SIC codes between
6000-6999). We exclude shelf offerings. We require that the offer price for the
issue must at least be $1 to prevent the possibility of bid-ask bounce dominating
our analysis. We also require that the ﬁnancial data on book value of total assets
(C OMPUSTAT data 6) and operating income (COMPUSTAT data 13) must be

available for the ﬁscal year of the SEO. Finally, for our operating performance

35

analysis, an SEO will be included in our sample only if the issuing ﬁrm does not
have a SEO in the last three years prior to the current SEO. Thus, once the ﬁrm has
a seasoned equity offering, the ﬁrm cannot re-enter the SEO sample within three
years of the issue date.

We then merge the resultant sample with the sample on board classiﬁcation
of Faleye (2007) for the period from 1995 to 2002.2 Falaye (2007) sample is
constructed from proxy statements ﬁled with the US Securities and Exchange
Commission in 1995. The sample excludes mutual funds, real estate investment
trusts, limited partnerships subsidiaries, and ﬁrms with incomplete data in
COMPUSTAT. The sample includes only those ﬁrms that maintain a unitary or a
classiﬁed board structure for the whole period from 1996 through 2002.
Furthermore, all ﬁrms included in his sample do not change their board structure
since 1990 thus ensuring that any sampled ﬁrms not only has the same board
structure throughout the empirical window of the study but also has not changed its
structure for at least ﬁve years prior to the sample period. This goes a long way in
mitigating the self-selection problem that has been well documented in the
literature regarding the endogenous relation between ﬁrm performance and board
structure. Since our operating performance analysis looks at a 3-year window
following a SEQ, we extend the Falaye sample till 2005. For ﬁrms which issue
stock during 2000-2002, we look at the issuing ﬁrm’s proxy statement to check
whether the ﬁrm maintains the same board structure, which it utilized prior to the
SEO, for a period of three years following the SEO. Merging our initial sample Of

SEOS with Faleye (2007) sample gives us a total of 210 SEOs.

 

3 We thank Olubunmi Faleye for generously providing us with his data.

36

Table 2.1 provides the total and yearly distribution of SEOS in our sample.
This table also reports the mean and median value of market capitalization (Market
Cap) at the end of the ﬁscal year prior to the issuance, the amount of money raised
through the SEO (Proceeds) and the amount of money raised through the SEO as a
fraction of the ﬁrm’s market capitalization (Oﬂer Size) at the end of the ﬁscal year
prior to the SEO. Panel A reports the summary for all 210 ﬁrms, Panel B for 108
ﬁrms with a classiﬁed board structure and Panel C for 102 ﬁrms with a non-
classiﬁed or unitary board structure. So we see that the sample is divided fairly
evenly between ﬁrms having classiﬁed boards and unitary boards and this
observation holds for all years in our sample. The maximum number of SEOs in a
year is 58 and this occurs in 1996 whereas the minimum number of SEOS in a year
is 5 and it occurs in 2001.

The mean (median) ﬁrm in our sample has Market Cap Of $870 million
($194 million). The mean (median) ﬁrm with classiﬁed boards has Market Cap of
$848 million ($254 million), while the mean (median) ﬁrm with unitary boards has
Market Cap of $894 million ($163 million). Also, the average ﬁrm in our sample
raised $96 million of proceeds in the offering and the proceeds represent 39% of
pre-issue market capitalization. The average ﬁrm having a classiﬁed board
structure raised $98 million and the proceeds represent 36% of pre-issue market
capitalization, whereas the average ﬁrm having a unitary board structure raised $93
million and the proceeds represent 42% of pre-issue market capitalization. This

large Offer Size indicates that the SEOS are important events for our sample ﬁrms.

37

Table 2.2 reports the industry distribution of SEOS in our sample. We use
the 12 industry classiﬁcation of Fama and French to classify the ﬁrms. The
classiﬁcation is obtained from Kenneth French’s web page. We ﬁnd that almost all
industries are represented in our SEO sample and the distribution of industry
groups are similar for classiﬁed and unitary boards suggesting that our analysis is
unlikely to suffer from industry induced biases. Firms belonging to the Business
equipment industry represent roughly 24% of our sample whereas Manufacturing,
Shops, Healthcare and Energy represent roughly 23%, 12%, 11% and 9%
respectively of our sample. The statistics are roughly similar for classiﬁed and

unitary boards.

2.2.2 Variable deﬁnition

Our key explanatory variable in the analysis is a dummy variable Cboard that
equals one if the issuing ﬁrm has a classiﬁed board and zero otherwise. We draw
control variables from the previous literature. We include market-to-book ratio (MB)
to control for the growth opportunity of a ﬁrm. MB is computed as the market value
of the ﬁrm’s equity plus the book value of assets minus the book value of equity
divided by the book value of assets. AS Kim and Pumanandam (2006) argue,
growth ﬁrms have more proﬁtable projects and investors believe that these ﬁrms
are less likely to conduct value-destroying investments. We expect market-to-book
ratio to be positively related to announcement returns. We also include past returns
(Past Return). We measure past return as the ﬁrm’s raw buy-and-hold return over a

period Of one year prior to the SEO issue date. Previous studies provide competing

38

views on the effect of ﬁrms’ past returns (e.g. Kim and Pumanandam (2006)).
Firms’ past returns can be considered as a proxy for the availability of good
projects. However, ﬁrms’ past return can also be considered as a proxy for
overvaluation of stocks since ﬁrms are more likely to issue equity when their stocks
are overvalued. Therefore, we do not clearly predict the relation between ﬁrms’
past returns and announcement returns.

We control for ﬁrm size (Log (TA)), which is deﬁned as the natural
logarithm of the ﬁrms’ total assets. Larger ﬁrms are more likely to be under greater
scrutiny and to be followed more actively by analysts and ﬁnancial press. Therefore,
ﬁrm size tends to reduce the information asymmetry (Ferreira and Laux (2007)).
We expect a positive relation between ﬁrm Size and announcement returns. Smith
(1977) posits that offer Size is positively related to announcement returns.
Following Kim and Pumanandam (2006), we include a ﬁrm’s cash divided by total
assets (Cash) to control for ﬁnancial slack of a ﬁrm since ﬁnancial slack of the ﬁrm
may reduce the problem of adverse selection.

We control for ﬁrm leverage (Leverage), which is deﬁned by the sum of
debt in current liabilities plus long term debt divided by total assets. Prior research
provides competing arguments on the effect of ﬁrm leverages. Leverage can reduce
the agency problems by controlling managerial discretion (Stulz (1990)). On the
contrary, a conﬂicting view suggests that high leverage can create incentives to
take risky negative-NPV projects at the expense of lenders (Jensen and Meckling
( 1976)). Therefore, we do not have a clear prediction about the relationship

between leverage and announcement reruns. Following Bates (2005), we include

39

CAPEX, which is deﬁned as the difference in capital investments as a fraction of
total assets for the issuing ﬁrm and the median ﬁrm in the same industry (based on
two-digit SIC code) as the issuing ﬁrm for the ﬁscal year following the issue. The
variable CAPEX gives us a measure of relative investment of the issuing ﬁrm
compared to the median ﬁrm in the same industry. Bates (2005) argues that this
measure of capital expenditure should, under rational expectations, provide a
reasonable ex-post proxy for ex-ante expected investment.

To control for the economy of scale effect (Ferreira and Laux (2007)), we
include Oﬂer Size, which is deﬁned as the amount of proceeds raised from the SEO
divided by market capitalization at the end of the ﬁscal year prior to the issue. We
expect the offer size of the equity issuance to be positively associated with
announcement returns. We create a dummy variable Secondary that equals one if
some secondary shares are included in the SEO offering and zero otherwise. Bates
(2005) shows that retention of proceeds from assets sales can lead to reduction in
shareholder welfare if the ﬁrm has poor growth opportunities. We include a dummy
variable Retention that equals one if the ﬁrm intends to retain the proceeds from the
SEO for any corporate purpose other than to retire debt or repurchase equity. If
more than one use of the proceeds is stated, we choose the ﬁrst stated use of
proceeds. This information is obtained by initially conducting a Factiva search. If
no information is found regarding the issuing ﬁrms intended use of proceeds, we
use the primary use of proceeds data provided in the SDC database. Since earnings

management may affect investors’ response to the announcement of SEO. we

40

control for Accrual, which is measured as the difference between net income and
cash flow from operations divided by total assets (Kim and Pumanandam (2006)).

Table 2.3 reports the mean (median) value for the control variables for the
whole sample and the sub-samples corresponding to classiﬁed and unitary boards.
The mean (median) value of Log (TA) for ﬁrms with classiﬁed boards is 5.508
(5.256), whereas the mean (median) value of Log (TA) for ﬁrms with unitary
boards is 5.018 (4.783). The statistics indicate that mean value of MB is 2.1 14 for
ﬁrms with classiﬁed boards and 2.657 for ﬁrms with unitary boards. The average
ﬁrm with classiﬁed boards has Leverage of 27.3%, while the average ﬁrm with
unitary boards has Leverage of 26.8%, respectively. The mean Cash is 0.101 for
ﬁrms with classiﬁed boards and 0.160 for ﬁrms with unitary boards.

Interestingly, ﬁrms with unitary boards are more likely to retain the
proceeds raised in the SEOS and spend more capital expenditures following the
SEOS. On average, 51.9% of the ﬁrms with classiﬁed boards indicate that the
proceeds raised in the SEOS will be retained for any corporate purpose, while
55.9 % of ﬁrm with unitary boards indicate that the proceeds will be retained. Also.
average CAPEX following the SEOS is 0.026 for ﬁrms with classiﬁed boards.
whereas average CAPEX following the SEOs is 0.033 for ﬁrms with unitary boards.
respectively. Therefore, if investors react more unfavorably to SEO announcements
by ﬁrms with classiﬁed boards than by ﬁrms with unitary boards, this is not
because investors worry about the absolute amount of the spending itself but
because they worry about the agency problems in the use of raised funds in ﬁrms

with classiﬁed board compared to ﬁrms with unitary board.

41

2.3 Empirical Tests

2.3.1 Announcement Returns

To get the announcement date of the SEO, we ﬁrst search for any indication of a
ﬁrm’s plan of an SEC on all publications included in Factiva including Wall Street
Journal and the Dow Jones News Retrieval Services. If we ﬁnd any news
publication of a ﬁrm’s plan of an SEO, we consider the ﬁrst occurrence of such a
publication as the announcement date provided it is earlier than the ﬁling date
provided in the SDC database. Otherwise, the ﬁling date is used as the
announcement date of the SEO.

Table 2.4 reports cumulative abnormal returns (CARS) around the SEO
announcement dates. Abnormal returns are computed using a standard market
model with parameters estimated over days -249 and -50 relative to the
announcement date. Panel A Shows CARS for the event window (-1, +1) days and
Panel B Shows CARS for the event window (-1, 0) days. For each event window,
the table reports the mean and median CAR for all SEOS in our sample, the sample
with only classiﬁed boards, the sample with only unitary boards, the test of
differences and their associated p-values based on t-tests for difference in means
and Wilcoxon rank-sum test for difference in medians.

The results in Table 2.4 Show that the average cumulative abnormal returns
(CARS) around SEO announcement for all SEOS is -2.2% over the announcement
window (-1, +1) and -1.1% over the announcement window (-1, 0). The results

suggest that on average, investors react negatively to announcements of SEOS,

42

which is consistent with the well documented previous literature on the negative
reaction to SEO announcements. The mean and median return differences between
unitary and classiﬁed boards are positive and statistically signiﬁcant for both event
windows. Speciﬁcally, on average, ﬁrms having a unitary board structure have
CARS of -1.4% and -0.2% over the announcement window (-1, +1) and (-1, 0),
respectively, while ﬁrms with a classiﬁed board structure have CARS of -3% and -
1.9% over the announcement window (-1, +1) and (-1, 0), respectively. Thus, on
average, classiﬁed boards earn 1.6% and 1.7% lower CARS over announcement
windows (-1, +1) and (-1, 0) respectively compared to unitary boards. Thus, the
results suggest that the market reacts more negatively to SEOS for ﬁrms that have a
classiﬁed board structure relative to SEOS for ﬁrms that have a non-classiﬁed or
unitary board structure. This result is consistent with our entrenchment hypothesis
that poor governance in the form of classiﬁed boards can partly explain the

negative reaction to SEOs.

2.3.2 Multivariate analysis of announcement returns
Table 2.5 presents the regression analysis of cumulative abnormal returns (CARS)
on board classiﬁcation. The dependent variable is the CARS for SEO ﬁrms during
the event window (-1, 1) days surrounding the announcement date. Qualitatively
similar results are obtained for the (-1, 0) event window.

In Model 1, the coefﬁcient on Cboard is negative and signiﬁcant at 5%
level of signiﬁcance, which indicates that market reacts more negatively to SEOS

by ﬁrms having classiﬁed boards relative to ﬁrms having unitary boards. This

43

suggests that outside investors worry more about the misuse of funds raised in the
SEO for ﬁrms having a classiﬁed board structure compared to ﬁrms having a
unitary board structure. Therefore, this result is consistent with our entrenchment
hypothesis.

The estimated coefﬁcient on the market-to—book ratio (MB) is positive and
statistically signiﬁcant at 1% level. The result indicates that investors believe that
ﬁrms with high growth opportunities conduct value-increasing investments by
raising ﬁinds ﬁom the SEOS. The coefficient on Past Return is negative and
statiStically signiﬁcant at 5% level, which supports the view that ﬁrms are more
likely to issues equities when their stocks are overvalued. The estimated coefﬁcient
of Log (TA) is positive and statistically signiﬁcant at 5% level. This evidence is
consistent with the argument that ﬁrm size tends to reduce the information
asymmetry. The coefﬁcients on other control variables in Model 1 are not
statistically different from zero. In Model 2, we add two more control variables
Retention and Accrual. The results are similar to those reported in Model 1 and two
added variables turn out to be insigniﬁcant.

In Model 3, we add the explanatory variable CAPEX and the interaction
term between CAPEX and Cboard to investigate the effect of the investors’
expectations of the increased capital expenditures after the SEC on the
announcement returns. The estimated coefﬁcient on the interaction term between
CAPEX and Cboard is negative and statistically signiﬁcant at 5% level. The result
support the view that the market believes that ﬁrms having a classiﬁed board

structure are more likely to conduct value-destroying expenditures using the

44

proceeds from the SEO relative to ﬁrms having a unitary board structure. The
coefﬁcient on Cboard is negative but loses its statistical signiﬁcance suggesting
that the potential to value-destroying expenditures could be the primary concern to

investors when ﬁrms having classiﬁed boards announce the SEO.

2.3.3 Multivariate analysis of change in operating performance

This section investigates the operating performance following the SEOS. We use a
matching ﬁrm benchmark to compute abnormal operating performance. Similar to
Barber and Lyon (1996) and Loughran and Ritter (1997), nonissuing matching
ﬁrms are selected on the basis of industry, asset size and operating performance.
Each issuing ﬁrm is matched with a ﬁrm that has not issued equity during the three
years prior to the issue date. To be included in the pool of candidate matching ﬁrms,
the ﬁrms should be listed in COMPUSTAT and report operating income (data 13)
and total assets (data 6) in a given calendar year. To be consistent with our sample
selection criteria, we exclude utilities and ﬁnancial ﬁrms and we include the ﬁrms
for US. common stocks listed on NYSE, NASDAQ, or AMEX. From this pool of
candidate matching ﬁrms, we identify those ﬁrms with the same historical two-
digit SIC code as the issuing ﬁrm and asset Size between 25% and 200% of the
issuing ﬁrm at the ﬁscal year of the issue (year 0). From these ﬁrms we select the
ﬁrm with the closest OIBD/assets ratio to that of the issuing ﬁrm as the matching
ﬁrm. Whenever no matching ﬁrm is available that meets these conditions, we
match on size and performance in the following way. We identify those ﬁrms with

asset size between 90% and 110% of the issuing ﬁrm at year 0. From these ﬁrms

45

we select the ﬁrm with the closest OIBD/assets ratio to that of the issuing ﬁrm as
the matching ﬁrm. If a matching ﬁrm issues equity during the three years following
the issue date, we do not replace it. However, if a matching ﬁrm is delisted, we
replace it with the next best matching ﬁrms as of the original matching date. We
require that the replacement ﬁrm must not have issued equity between the issue
date and the replacement date. If an issuing ﬁrm is delisted, the matching ﬁrm is
also removed at the same time. Therefore, we use the same number of event years
for the issuing ﬁrm and matching ﬁrm.

Table 2.6 reports the result of the regression analysis of change in matched-
ﬁrm-adjusted operating performance. The dependent variable is the change in
matched-ﬁrm-adjusted ROA, measured as the difference between median of
matched-ﬁrm-adjusted ROA for the three years following the issue (year 1, 2 and 3)
and matched-ﬁrm-adjusted ROA in the prior year to the issue (year -1). The
matched-ﬁrm-adjusted ROA is calculated as the difference between the issuing
ﬁnns’s ROA and the corresponding ROA of the matching ﬁrm. We ﬁnd that the
coefﬁcient on C board is negative and statistically Signiﬁcant at 5% level. Firms
having a classiﬁed board structure have 4.7 % lower abnormal operating
performance compared to ﬁrms having a unitary board structure. The result
suggests that ﬁrms with classiﬁed boards suffer from agency problems when the
ﬁrms raise funds through the SEOS. This evidence supports our entrenchment

hypothesis.

2.4 Conclusion

46

This paper attempts to understand the relevance of board classiﬁcation in
explaining the well documented negative market reaction to seasoned equity
offerings. In particular, we show that announcement returns of SEOs are more
negative for ﬁrms having a classiﬁed board structure compared to ﬁrms having a
unitary board. More importantly, we Show that the market believes that ﬁrms
having a classiﬁed board structure are more likely to misuse the proceeds from the
SEOS relative to ﬁrms having a unitary board structure. Our analysis of change in
matched-ﬁrm-adjusted operating performance Show that ﬁrms with classiﬁed
boards earn signiﬁcantly lower abnormal operating performance following SEOS
compared to ﬁrms with unitary boards. This is consistent with the market’s
expectation that ﬁrms with classiﬁed boards suffer from agency problems when the
ﬁrms raise funds through the SEOs. Overall, our results support the entrenchment
hypothesis that classiﬁed boards entrench managers and are ineffective in

preventing them from misuse of funds raised in SEOS.

47

Chapter 2 Appendix

48

SEO sample distribution by issue year

Table 2.1

This table reports the distribution of our sample of seasoned equity issues (SEOs) by issue year.
Panel A reports mean and median values for all SEOs; Panel B and C reports yearly distribution by
dividing the original sample into classiﬁed boards and unitary boards respectively. Market Cap
denotes the market capitalization of the issuing ﬁrm at the end of the ﬁscal year prior to the security
issuance, in millions of dollars. Proceeds represents the money amount raised through the SEOS
measured in millions of dollars. Oﬂer Size represents the proceeds raised through the SEO as a
fraction of its market capitalization at the end of ﬁscal year prior to the security issuance.

Panel A: All SEOs

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Mean Median
Year yfugggg Mgglget Proceeds Offer Size Market Cap Proceeds Offer Size
1995 50 933 67 0.31 148 45 0.25
1996 58 364 86 0.49 153 58 0.36
1997 31 200 53 0.42 134 41 0.28
1998 15 2192 180 0.37 252 57 0.25
1999 17 1644 116 0.46 256 77 0.23
2000 16 1269 177 0.45 502 l 18 0.20
2001 5 690 75 0.13 639 58 0.15
2002 18 1349 124 0.21 523 94 0.16
Total 210 870 96 0.39 194 58 0.26
Panel B: Classiﬁed boards
Mean Median
Year Eggs; Méglget Proceeds Offer Size Market Cap Proceeds Offer Size
1995 26 332 62 0.28 188 54 0.26
1996 29 379 92 0.49 195 67 0.37
1997 18 206 50 0.42 155 44 0.27
1998 6 2783 201 0.23 267 60 0.22
1999 7 3288 154 0.24 294 105 0.23
2000 8 700 161 0.55 719 139 0.40
2001 3 430 65 0.15 389 58 0.15
2002 1 1 1970 148 0.17 896 106 0.10
Total 108 848 98 0.36 254 62 0.25
Panel C: Unitary boards
Mean Median
Year (Nfugébgg Méglget Proceeds Offer Size Market Cap Proceeds Offer Size
1995 24 1583 72 0.33 135 36 0.24
1996 29 349 81 0.48 134 44 0.32
1997 13 191 57 0.40 118 33 0.33
1998 9 1798 165 0.47 132 50 0.43
1999 10 494 89 0.62 206 70 0.26
2000 8 1837 194 0.35 325 112 0.15
2001 2 1079 90 0.09 1079 90 0.09
2002 7 371 85 0.27 464 91 0.25
Total 102 894 93 0.42 163 54 0.27

 

49

Table 2.2
Industry distribution

This table reports the industry distribution of seasoned equity issues (SEOS) in our sample. The
SEO ﬁrms are classiﬁed into industries based on the 12 industry classiﬁcation by Fama and
French.

 

 

 

Classiﬁed Unita
Industry All SEOS boards boar (I:
Business equipment 50 26 24
Chemicals 5 3 2
Consumer durables 8 2 6
Consumer nondurables 9 5 4
Energy 19 13 6
Healthcare 23 10 13
Manufacturing 48 30 18
Shops 26 11 15
Telecom 6 2 4
Other 16 6 10
Total 210 108 102

 

50

Table 2.3
Summary statistics

This table reports the mean (median) value for control variables used in the regression analyses. Log
(TA) denotes the log of total assets (data 6) in the ﬁscal year prior to the issue. CAPEX is the difference
in capital investments as a fraction of total assets (data 128/data 6) for the issuing ﬁrm and the median
ﬁrm in the same industry (based on two-digit SIC code) as the issuing ﬁrm for the ﬁscal year following
the issue. MB is deﬁned by the market value of assets divided by book value of assets, which is
calculated as (data 6 + data 199*data 25 - data 60)/data 6. Leverage is the ratio of the sum of debt in
current liabilities (data34) plus long term debt (data 9) to total assets (data6) in the ﬁscal year prior to the
issue. Cash denotes cash (data 1) as a fraction of total assets (data 6) in the ﬁscal year prior to the issue.
Past Return is the raw buy-and-hold return over a period of one year prior to the SEO issue date. Offer
Size represents the proceeds raised through the SEO as a fraction of its market capitalization at the end
of ﬁscal year prior to the security issuance. Accrual is the ratio of the difference between cash ﬂow from
operating activities and net income to total assets in the ﬁscal year prior to the issue, which is calculated
as (data 172 — data 308) / data 6. Secondary is a dummy variable that takes the value of one if some
secondary shares are included in the SEO offering and zero otherwise.

 

 

All Classiﬁed Board Unitary Board
(N=210) (N=108) (N=102)
Log (TA) 5.270 5.508 5.018
(5.078) (5.256) (4.783)
CAPEX 0.029 0.026 0.033
(0.004) (0.005) (0.003)
MB 2.378 2.114 2.657
(1.574) (1.506) (1.761)
Leverage 0.270 0.273 0.268
(0.253) (0.280) (0.225)
Cash 0.129 0.101 0.160
(0.054) (0.039) (0.057)
Past Return 1.175 1.071 1.286
(0.758) (0.669) (0.884)
Offer Size 0.388 0.361 0.417
(0.261) (0.253) (0.268)
Accrual -0.050 -0.045 -0.055
(-0.044) (-0.040) (-0.052)
Retention(dummy) 0.538 0.519 0.559
(1.000) (1.000) ( 1.000)
Secondary(dummy) 0.405 0.398 0.412
(0.000) (0.000) (0.000)

 

51

Table 2.4
Announcement abnormal returns and board classiﬁcations

This table reports cumulative abnormal returns (CARS) around the SEO
announcement dates. Abnormal returns are computed using a standard market model
with parameters estimated over days -249 and -50 relative to the announcement date.
Panel A shows CARS for the event window (-1, +1) days and Panel B shows CARS for
the event window (-1, 0) days. P-values in parentheses are based on t-tests for
difference in means and Wilcoxon rank-sum test for difference in medians.

Panel A : Event window: (-1, +1)

 

 

 

 

 

Mean Median N
All SEOs -0.022 -0.028 210
Unitary Boards (A) -0.014 -0.020 102
Classiﬁed Boards (B) -0.030 -0.033 108
(A) - (B) 0.016 0.015
(0.095) (0.084)
Panel B : Event window: (—1, 0)
Mean Median N
All SEOs -0.011 -0.015 210
Unitary Boards (A) -0.002 -0.006 102
Classiﬁed Boards (B) -0.019 -0.019 108
(A) - (B) 0.017 0.013
(0.030) (0.029)

 

52

Table 2.5
Regression analysis of announcement returns on board classiﬁcation

This table reports the results of the regression analysis of cumulative abnormal returns (CARS) on
board classiﬁcation. The dependent variable is the CARS for SEO ﬁrms during the event window (-1,
1) days surrounding the announcement date. Cboard is a dummy variable that equals one if the issuing
ﬁrm has a classiﬁed board, zero otherwise. MB is deﬁned by the market value of assets divided by
book value of assets, which is calculated as data (6 + data 199*data 25 - data 60)/data 6. Past Return is
the raw buy-and-hold return over a period of one year prior to the issue date. Log (TA) denotes the log
of total assets (data 6) in the ﬁscal year prior to the issue. Oﬂer Size represents the proceeds raised
through the SEO as a fraction of its market capitalization at the end of the ﬁscal year prior to the
security issuance. Leverage is the ratio of the sum of debt in current liabilities (data 34) plus long term
debt (data 9) to total assets (data 6) in the ﬁscal year prior to the issue. Cash denotes cash (data 1) as a
fraction of total assets (data 6) in the ﬁscal year prior to the issue. Secondary is a dummy variable that
takes the value of one if some secondary shares are included in the SEO offering and zero otherwise.
Retention is a dummy variable that equals one if the ﬁrm intends to retain the proceeds raised in the
SEO and zero otherwise. Accrual is the ratio of the difference between cash ﬂow from operating
activities and net income to total assets in the ﬁscal year prior to the issue, which is calculated as (data
172 — data 308) / data 6. CAPEX is the difference in capital investments as a fraction of total assets
(data 128/data 6) for the issuing ﬁrm and the median ﬁrm in the same industry (based on two-digit SIC
code) as the issuing ﬁrm for the ﬁscal year following the issue. The coefﬁcients on the intercepts are
not reported. All models include industry dummies (based on two-digit SIC code) and year dummies,
whose coefﬁcient estimates are also not reported. P-values in parenthesis are corrected for
heteroscedasticity.

 

 

 

Model 1 Model 2 Model 3

Estimate p-value Estimate p-value Estimate p-value
Cboard -0.021 0.062 -0.022 0.054 -0.019 0.127
MB 0.006 0.004 0.006 0.006 0.007 0.004
Past Return -0.010 0.047 -0.010 0.046 -0.009 0.132
Log (TA) 0.010 0.013 0.009 0.036 0.007 0.100
Offer Size 0.023 0.288 0.016 0.417 0.019 0.367
Leverage -0.001 0.976 0.009 0.769 0.009 0.770
Cash 0.039 0.358 0.031 0.447 0.008 0.878
Secondary 0.017 0.209 0.018 0.185 0.019 0.187
Retention 0.011 0.341 0.011 0.358
Accrual -0.077 0.216 -0.064 0.373
CAPEX 0.071 0.329
Cboard x CAPEX -0.246 0.026
N 210 210 199
R-Squared 0.267 0.278 0.299

 

53

Table 2.6
Regression analysis of change in operating performance on board
classiﬁcation

This table reports the result of the regression analysis of change in matched-ﬁrm-adjusted
operating performance on board classiﬁcation. The dependent variable is the change in
matched-ﬁrm-adjusted ROA, measured as the difference between the median of matched-
ﬁrm-adjusted ROA for the three years following the issue (year 1, 2 and 3) and matched-
ﬁrrn-adjusted ROA in the prior year to the issue (year -1). Cboard is a dummy variable that
equals one if the issuing ﬁrm has a classiﬁed board, zero otherwise. MB is deﬁned by the
market value of assets divided by book value of assets, which is calculated as (data 6 + data
199*data 25 - data 60)/data 6. Log (TA) denotes the log of total assets (data 6) in the ﬁscal
year prior to the issue. Offer Size represents the proceeds raised through the SEO as a fraction
of its market capitalization at the end of the ﬁscal year prior to the security issuance. The
coefﬁcients on the intercepts are not reported. P-values are adjusted for heteroscedasticity.

 

 

 

Estimate p-Value
Cboard -0.047 0.041
MB 0.007 0.258
Log(TA) 0.004 0.635
Offer Size -0.009 0.645
N 202
R-Squared 0.031

 

54

Chapter 3

Product Market Competition and Firms’ Credit Ratings

3.1 Introduction

It has long been argued that product market competition reduces managerial slack
and improves efﬁciency. Speciﬁcally, managers in competitive industries are under
continuous pressure to maximize ﬁrm value to ensure the survival of ﬁrm and thus
themselves (e. g. Alchian (1950) and Friedman (1953)). Empirical studies show that
product market competition helps align the interests of managers and shareholders
and thus results in improved efﬁciency. For example, Grullon and Michaely (2007)
argue that product market competition discipline managers to payout excess cash.
They Show a negative relation between corporate payout and the level of industry
concentration. Grifﬁth (2001) ﬁnds that an increase in product market competition
leads to an increase in overall levels of efﬁciency and growth rates.

Allen and Gale (1999) argue that product market competition can be a more
effective governance mechanism than market for corporate control or oversight by
institutions. Governance mechanisms that discipline self-interested managers can
be beneﬁcial to both shareholders and bondholders. However, managers closely
aligned with shareholders may take actions that can transfer wealth from creditors
to shareholders (Jensen and Meckling (1976)). Speciﬁcally, managers who have
incentives to maximize shareholder value are more likely to invest in risky projects

at the expense of debtholders. Ortiz-Molina (2008) argues that managerial risk-

55

taking incentives are positively related to the size of managerial ownership stakes.
He ﬁnds a positive relationship between managerial ownership and borrowing costs.

In this paper, we investigate whether the link between product market
competition and managerial incentives affects ﬁrms’ credit ratings. Product market
competition can discipline managers to maximize shareholder value for their job
security. Managerial incentives to maximize shareholder value can lead to
undertaking more risky projects, which lowers ﬁrm’s credit rating. Therefore we
expect that an increase in product market competition leads to higher ﬁrms’ credit risk
and lower credit ratings.

Using a data set of 6589 ﬁrm-year observations on 1155 US. ﬁrms for the
years 1996- 2006, we empirically investigate whether the link between managerial
incentives and product market competition have different effects on shareholders
and creditors. We employ Herﬁndahl-Hirschman Index (HHI) based on
COMPUSTAT as the proxy for product market competition. As hypothesized, we
ﬁnd that market competition is negatively related to ﬁrms’ credit ratings. This
result holds even after controlling for other factors such as ﬁrm characteristics and
governance attributes that have been documented in the literature to affect ﬁrms’
credit ratings. We also ﬁnd that market competition is positively related to
shareholder value measured as adjusted Tobin’s Q, and ﬁrms in more competitive
markets tend to make more R&D and advertising expenditures. Our results suggest
that managers are more likely to be aligned with shareholders in more competitive
markets and thus undertake riskier investments to maximize shareholder value,

which lowers ﬁrms’ credit ratings.

56

Our study extends the work by Ortiz-Molina (2008). He argues that
managerial incentives to maximize Shareholder value can be detrimental to
bondholders and prospective bondholders anticipate managers’ future choice of risk
level at the issue of debt. He ﬁnds that at-issue yield spreads of corporate bonds are
positively related to managerial ownership. Our ﬁndings indicate that product
market competition can create managerial incentives to maximize shareholder
value and thus are detrimental to creditors. Our study also extends the literature that
investigate the effects of corporate governance on credit ratings (e.g. Ashbaugh-
Skaife, Collins, and LaFond (2006) and Bradley, Chen, Dallas and Snyderwine
(2007)) by providing evidence that product market competition can attenuate the
agency conﬂict between managers and shareholders as an effective governance
mechanism but have negative effect on credit ratings.

There is a growing literature about the effectiveness of product market
competition as a corporate governance mechanism. For example, Grullon and
Michaely (2007) argue that product market competition discipline managers to
disgorge excess cash. They Show that ﬁrms in more competitive markets have
signiﬁcantly higher payout ratios than ﬁrms in less competitive markets. Xiumin
Martin (2007) argues that product market competition reduces managers‘
tendencies to waste cash and thus increases ﬁrm value. The results in this paper add
to the literature by suggesting that product market competition can effectively
discipline managers to maximize shareholder value and thus create managerial

incentives to take risky investments.

57

The remainder of the paper is organized as follows. Section 2 describes the
sample construction procedure, deﬁnes the variables employed, and presents
descriptive statistics. Section 3 investigates the relation between product market
competition and Tobin’s Q. Section 4 explores the relation between product market
competition and credit ratings. Section 5 examines the link between product market
competition and risky investments. Section 6 concludes.

3.2 Sample construction

To construct our sample, we use ﬁve databases in this study: the Compustat
Industrial Annual (COMPUSTAT) database, the Center for Research in Security
Prices (CRSP) database, the Executive Compensation (ExecuComp) database, the
Investor Responsibility Research Center (IRRC) database and the Thomson Reuters
(Thomson) database.

We use IRRC database to collect governance index (G-Index) score.
Following Gompers et a1. (2003), we assume that for the years between two
consecutive reports, G-Index score is the same as in the previous report year. Also,
board information iS from the IRRC data for the years from 1996 through 2006. We
use ExecuComp to collect the stock option holdings, stock holdings, and
compensation information for the top ﬁve executives. The Thomson data is used to
collect information about institutional holdings of ﬁrms. Finally, we use
C OMPUSTAT to construct ﬁrm-Speciﬁc ﬁnancial characteristics and use C RSP to
obtain daily and monthly stock returns.

To be included as a ﬁrm-year observation in this study, the following

conditions must be satisﬁed:

58

(1) The G-Index score and board information Should be available in the
IRRC dataset.

(2) Data on stock and stock option holdings for top 5 executives should also
be available from the ExecuComp.

(3) Information on institutional holdings should be available from Thomson
data.

(4) Firm-speciﬁc ﬁnancial characteristics and returns data should be
available in the COMPUSTAT and CRSP, respectively.

By merging and requiring these conditions, we have a data set of 65 89 ﬁnn-

year observations on 1155 ﬁrms for the years 1996 through 2006.

3.3 Variable Deﬁnitions

3.3.1 Measuring product market competition

We use Herﬁndal-Hirschman index (HHI) is as a measure of product market
completion in this study. Following Giroud and Mueller (2008), HHI index is

deﬁned as the sum of squared market shares,

HHIj, = ”I s

2
i=1 ijt

where 5,), is the market share of ﬁrm i in industry j in year t and N,- is the total
number of ﬁrms in industry j in year t. Market Shares are based on ﬁrms’ sales.
Industry is classiﬁed by three-digit SIC codes. When we compute HHI index, we
use all ﬁrm year Observations in the Compustat unless sales are missing or negative.

Low values of HHI index indicate that a lot of competing ﬁrms Share the market.

59

On the other hand, high values of the HHI indicate that a small number of ﬁrms has

concentrated market shares.

3.3.2 Variables for Tobin’s Q regression

We measure the wealth of shareholders using Tobin’s Q. Similar to Gompers, Ishii
and Metrick (2003) and Giroud and Mueller (2008), industry adjusted Tobin’s Q is
computed as the difference between the ﬁrm’s Q and the median Q in the same
industry as deﬁned by three-digit SIC code. Q is deﬁned by the market value of
assets divided by book value of assets.

To isolate the effect of market competition on the wealth of shareholders,
we control for other variables which are known to have an effect on Tobin’s Q. We
draw control variables from the previous literature (see e.g. Gompers, Ishii and
Metrick (2006) and Faleye (2007)). There are two groups of control variables. The
ﬁrst group of the variables is related to ﬁrm characteristics. We control for ﬁrm
size (Size), which is deﬁned as the natural logarithm of the ﬁrms’ total assets. We
measure ﬁrm leverage (Leverage) as the sum of debt in current liabilities plus long
term debt divided by total assets. Gompers, Ishii and Metrick (2006) point out
younger ﬁrms are more likely to have high Tobin’s Q since they usually have more
grth opportunities. To control for the effect of ﬁrm age (Log (firm age)) on Q,
we include the logarithm of the number of years Since the ﬁrm has ﬁrst appeared on
CRSP. F inn proﬁtability (ROA) is net income before extraordinary items divided
by total assets. If a ﬁrm is included in S&P 500 index, this would cause a higher

Tobin’s Q as Morck and Yang (2001) Show. Therefore, we create a dummy

60

variable for S&P 500 index inclusion (S&P 500 Index Dummy), which is equal to
one if the ﬁrm is included in S&P 500 index and zero otherwise.

The second group of the variables is related to governance attributes. We
include managerial ownership (Insider Ownership), which is deﬁned as the
aggregate percentage ownership of common stocks held by top ﬁve executives.
Stock holdings by institutional investors (Institutional Ownership) are computed as
aggregate percentage ownership of shares by institutional investors which own at
least 5% of outstanding shares. McConnell and Servaes (1990) Show that
institutional ownership is positively associated with Tobin’s Q. However, the
argument that institutional investors help to align the interests of shareholders and
managers is contested (Shleifer and Vishny (1997)).

Board independence (Independent director) is computed as the percentage
of independent directors on the board. Prior studies Show mixed results for the
relation between board independence and ﬁrm performance. For example, Bhagat
and Black (2000) Show that board independence has no relation with Tobin’s Q
unlike generally accepted argument of positive relation between them. We deﬁne
the level of equity-based compensation (Equity Compensation) as the sum of the
Black-Scholes value of new stock options received by the top ﬁve executives as a
percentage of total compensation paid to them. Total compensation is comprised of
salary, bonus, other annual compensation, value of restricted stock granted, Black-
Scholes value of new stock Options granted, long-terrn incentive payouts, and all
other compensations. We expect a positive effect of Equity Compensation on

Tobin’s Q. since more equity-based compensation would incentivize manager to

61

take more risky behavior and would be more likely to align manager interest and
Shareholder interest.

We include governance index (G—Index) developed by Gompers, Ishii and
Metrick (2003). Lower G-lndex values indicate strong shareholder rights based on
antitakeover provisions. Thus we expect a negative relation between G-Index value
and Tobin’s Q. As Daines (2002) shows, Delaware incorporation may have a
positive impact on Tobin’s Q. However, Subramanian (2004) provide evidence that
small ﬁrms drive the Delaware effect. We include a dummy variable for Delaware
incorporation (Delaware Dummy), which is equal to one if the ﬁrm is incorporated

in Delaware and zero otherwise.

3.3.3 Variables for ﬁrms’ credit rating regression

Firms’ credit ratings (Credit Rating) are obtained from S&P’s long-term issuer
credit ratings in Compustat (data280). S&P reports ratings based on its assessment
of the creditworthiness of a ﬁrm as a senior debt obligor. Following Ashbaugh-
Skaife, Collins, and LaFond (2006), S&P ratings are classiﬁed into 7 categories as
shown in Table 3.1.

Based on the previous studies, we control for a number of ﬁrm
characteristics and governance attributes besides HHI index. The ﬁrst group of
variables is ﬁrm characteristics. We control for ﬁrm size (Size). Ortiz-Molina (2008)
argues that investments in larger ﬁrms are safer than investments in smaller ﬁrms
because larger ﬁrms have larger asset base as collateral. Therefore we expect a

positive relation between ﬁrm size and credit rating. We also control for ﬁrm

62

volatility (Volatility), which is deﬁned as standard deviation of monthly stock
returns over the past 5 years.

Following Ashbaugh-Skaife, Collins and LaFond (2006), we include
several variables to control for default risk. These variables are interest coverage,
ROA, leverage, and Loss Dummy. Interest Coverage is calculated as operating
income before depreciation divided by interest expense. We expect Interest
Coverage to be positively related to credit ratings. Unproﬁtable ﬁrms or highly
leveraged ﬁrms are more likely to default. Therefore we expect that credit ratings
are positively associated with ROA and negatively associated with Leverage. Loss
Dummy is equal to one if ROA is negative in the current and prior ﬁscal year, zero
otherwise. Loss Dummy is expected to be negatively related to credit ratings since
unproﬁtable ﬁrms are more likely to default.

Ashbaugh-Skaife, Collins and LaFond (2006) argue that a ﬁrm with
subordinated debt is more risky because of the different claims to assets by
creditors. Subordinated Debt Dummy is equal to one if the ﬁrm has subordinated
debt and zero otherwise. Firrn’s capital intensity (Capital Intensity) is computed as
net property, plant, and equipment divided by total assets. We expect ﬁrm’s capital
intensity to be positively associated with credit ratings since greater capital
intensity indicates less risk to creditors.

The second group of variables is governance attributes. Ortiz-Molina (2008)
posits that a higher level of managerial Stock and stock Option holdings give
managers risk-taking incentives and thus result in a closer alignment of manager

and shareholder interests. The author ﬁnds that borrowing costs are positively

63

related to managerial stock and stock option holdings. Therefore, we expect
managerial ownership (Insider Ownership) to be positively related to credit ratings.
Alpha is the sum of the number of common stock, restricted stock and unexercised
stock options held by top ﬁve executives as a percentage of total shares outstanding.
Alpha is also expected to be positively related to credit ratings.

We control for stock holdings by institutional blockholders (Institutional
Ownership). Prior research provides competing arguments on the effect of
institutional investors on ﬁrm stakeholders (Bhojraj and Sengupta (2003)).
Institutional investors can play a role of efﬁcient monitoring, which is beneﬁcial to
both shareholders and bondholders. On the contrary, a conﬂicting view suggests
that institutional investors can exert improper inﬂuence over management for their
private beneﬁts that are harmful to minority shareholders and bondholders. In other
words, institutional investors inﬂuence managers to transfer wealth from
bondholders to shareholders. Therefore, we do not have a clear prediction about the
relationship between stock holdings by institutional investors and credit ratings.

We also control for a percentage of the independent directors on the board
(Independent Director). Bhojraj and Sengupta (2003) argue that ﬁrms with more
outside directors are more likely to monitor effectively management behaviors and
thus create beneﬁts to both shareholders and bondholders. They empirically Show a
negative relation between the fraction of non-ofﬁcer directors on the board and
borrowing costs. This leads us to expect a positive relation between Independent

Director and credit ratings.

64

We include governance index score (G—Index) developed by Gompers, Ishii,
and Metrick (2003). They posit G-Index represents the level of Shareholder rights
based on antitakeover provisions. The prior research provides competing views on
the effect of G—Index on bondholder. Managers are entrenched with strong
antitakeover provisions (high G-Index). This leads to increased opportunistic
behavior by managers, which is detrimental to both shareholders and bondholders.
However, a competing view suggests strong antitakeover provisions are beneﬁcial
to bondholders since it lowers the likelihood of takeovers, which can transfer
wealth from bondholders to Shareholders (Klock, Mansi and Maxwell (2004)).
Thus, we do not have clear prediction about the relationship between G-Index and
credit ratings.

We also include a dummy for Delaware inclusion (Delaware Dummy).
Francis, Hasan, Jonh and Waisman (2006) conjecture that cost of debt is higher for
ﬁrms incorporated in takeover friendly states like Delaware because of a wealth
transfer from bondholders to shareholders in the event of takeover. They show that
ﬁrms incorporated in Delaware actually have higher cost of debt. We predict
Delaware incorporation is negatively associated with credit ratings.

Following Bradley, Chen, Dallas and Snyderwine (2007), we control for a
quality of reported earnings (Earnings Quality). They Show a positive relation
between ﬁrm’s earnings quality and credit rating. To measure the quality Of

reported earnings, they estimate following Fama-French three-factor model

65

augmented with accrual quality (AQ) factor of Ecker, Francis, Kim, Olsson and
Schipper (2006)3 :

R, z (1 + ﬂjMKT '1' ﬂzSMB + ﬂ3HML + ,B4AQ

R, is ﬁrm’s daily excess stock returns. AQ is calculated as the standard deviation of
the residuals by estimating the regressions of the total current accruals on past,
current, and future cash ﬂow from operations, change in revenues, and gross value
of property, plant, and equipment. The higher values of loadings on AQ factor
indicate a lower quality of reported earnings Since it suggests a high variance in the
accruals’ estimation. For easy interpretation, we multiply the loadings on the AQ
factor by -l and deﬁne the resulting values as Earnings Quality. Therefore, high
value of Earnings Quality corresponds to high quality of reported earnings. We
expect Earnings Quality to be positively related to credit ratings.

Finally, we control for predation risk using the proxy variable for the
correlation of ﬁrm stock returns with its corresponding industry stock returns
(Return Correlation) following Haushalter, Klasa and Maxwell (2009). The authors
use the variable as a proxy for interdependence of investment opportunities. They
argue that there is a positive relation between a correlation of a ﬁrm’s stock returns
with its industry counterparts and the expected interdependence of its investment
opportunities. The variable is constructed as the estimated coefﬁcient on the
industry return index (1m) from the following regression of Haushalter, Klasa and

Maxwell (2009),

I.t : )6!) + ,BIIM + ,62 ind

 

" We are grateful to Frank Ecker for generously providing us with the accrual quality factor.

66

where r, is a ﬁrm’s monthly stock returns for our sample period, [M is the monthly
equally-weighted market returns , and 1,)“, is the monthly equally-weighted index
return for the same industry as deﬁned by two-digit SIC code. Following
Haushalter, Klasa and Maxwell (2009), we require that a ﬁrm has at least 24
months of returns to compute the correlation. Also, we calculate the index based on
two-digit SIC code to include enough number of ﬁrms in the index for a
meaningful index returns. We expect Return Correlation to be negatively related to

credit rating.

3.4 Descriptive Statistics
Table 3.2 provides the descriptive statistics for the variables in this study. Included
are the number of observations, mean, median, standard deviation, 25th percentile
and 75th percentile values for the variables in the analysis. The median value of
credit rating is 4.0, which corresponds to the range of BBB+ to BBB- in the S&P'S
rating scheme. On average, 72% of the sample has an investrnent-grade credit
rating. The mean (median) value of adjusted Tobin’s Q is 0.25 (0.03). HHI index in
the sample has a mean of 0.16 and a median of 0.1 1, with 25th and 75th percentile
values of 0.06 and 0.21, respectively. The mean value of Return Correlation is 0.82.
Turning to the variables of ﬁrm characteristics, total assets has a mean and
median of $17.14 billion and about $3.77 billion, respectively. Average ﬁrm age of
the sample is about 32 years. On average, 45% of the sample ﬁrms are included in
S&P 500 index. Therefore, ﬁrms in our sample are generally old and large, which

is consistent with the argument that ﬁrms issue public debt in later stage of their

67

life. The average ﬁrm in our sample has a ROA of 4%, a leverage ratio of 31% and
a volatility of 11%. The descriptive statistics indicates 6% of our sample ﬁrms have
negative ROA in the current and prior ﬁscal year and 16% of our sample ﬁrms
have subordinated debt. The mean interest coverage is 15.32, while the median
interest coverage is 6.37, which indicates the distribution is right-skewed (positive
skewness). We ﬁnd the mean and median capital intensity are 34% and 29%, with
the 25th and 75th percentiles equal to 15% and 53%, respectively.

The remaining variables are governance attributes. The table shows top ﬁve
executives own 1.69% on average. Institutional blockholders own, on average,
8.45% with a median of 6.69%. The average percentage of independent directors
on the board has a mean value of 67.66% and a median value of 70%. On average,
the value of 2.22% is the sum of common stock, restricted stock, and unexercised
stock options held by top ﬁve executives as a percentage of total shares outstanding.
The mean and median values of Earnings Quality are 0.07 and 0.10, respectively.
Recall that larger values of these variables reﬂect higher quality of reported
earnings. The G—Index has a mean of 9.76, a median of 10, a standard deviation of
2.6, and 25th and 75th percentile values of 8 and 12, respectively. On average, 57%
of our sample ﬁrms are incorporated in Delaware. The sum of the value of new
Stock options to the top 5 executives is, on average, 34.02% as a percentage of total
compensation, with the 25th and 75th percentiles equal to 14.78% and 51.19%,

respectively.

3.5 The Relation between Product Market Competition and Tobin’s Q

68

As Bradley, Chen, Dallas and Snyderwine argue, ﬁrm value (Tobin’s Q) is a more
direct concern for shareholders as compared with bondholders. To investigate the
effect of market competition on shareholder value, we employ the following model,

Adjusted Tobin ’s Q = f (HHI, ﬁrm characteristics, governance attributes)
The variables of ﬁrm characteristics include Size, ROA, Leverage, Log (firm age),
and S&P 500 Index Dummy. The variables of governance attributes include Insider
Ownership, Independent Director, Institutional Ownership, Equity Compensation,
G-Index, and Delaware Dummy. The deﬁnitions of the above variables are in the
Appendix. We estimate the above model using pooled OLS with year and industry
dummies. The p-values are based on the standard errors which are robust to
heteroskedasticity and adjusted for ﬁrm clustering.

Table 3.3 presents the regression results regarding the effect of market
competition on adjusted Tobin’s Q. Model 1 reports the results of regression of
adjusted Tobin’s Q on HHI index, together with the variables of ﬁrm characteristics.
Model 2 add the variables of governance attributes in addition to the variables in
Model 1. The results in Model 1 and Model 2 Show that the coefﬁcients on HHI
index are negative and statistically signiﬁcant at 5% level in both regressions. The
results indicate that managers are more aligned with shareholders in a more
competitive market and thus increase ﬁrm value. Moreover, the results are also
economically signiﬁcant. In Model 2, the coefﬁcient is -0.384. We ﬁnd that
moving from the 25th to the 75th percentiles of HHI index results in a decline of
5.76% in adjusted Tobin’s Q. Therefore, changes in market competition have a

considerable effect on the shareholder value.

69

In terms of variables of ﬁrm characteristics, the results from Model 2 show
that the coefﬁcient of Size is negative and statistically signiﬁcant, while the
coefﬁcient of ROA is positive and statistically signiﬁcant. As expected, a ﬁrm’s
inclusion in S&P 500 index has a positive effect on Tobin’s Q. This ﬁnding is
consistent with Morck and Yang (2001). The estimated coefﬁcients on Leverage
and Log (ﬁrm age) are not statistically different from zero.

Turning to the variables of governance attributes, the coefﬁcient estimate of
* Equity Compensation is negative and statistically Signiﬁcant at 1% level. This
indicates that high level of equity-based compensation make the managers more
closely aligned with shareholders, which leads to increase in ﬁrm value. The
coefﬁcient on G—Index is negative and signiﬁcant at 1% level. One standard
deviation increase in G-lndex results in 5.98% decrease in adjusted Tobin’s Q. This
evidence suggests that weak shareholder rights have a negative impact on ﬁrm
value. Interestingly, the coefﬁcient on Delaware Dummy is not statistically
different from zero. We surprisingly ﬁnd that the estimated coefﬁcient on
Institutional Owners/zip is negative and statistically signiﬁcant at 1% level, which
supports the view that institutional blockholders may exert their inﬂuence over
management for their private beneﬁts that are harmful to minority Shareholders
(Shleifer and Vishny (1997)). The coefﬁcients on Insider Ownership and

Independent Director are not statistically different from zero.

3.6 The Relation between Product Market Competition and Firms’ Credit

Ratings

70

3.6.1 Product Market Competition and Credit Ratings

We examine the effect of market competition on the credit ratings using the
following model,

Credit Ratings = f (HHI, Return Correlation, ﬁrm characteristics, governance
attributes)

Return Correlation is used as a proxy variable for Predation Risk. The included
variables of ﬁrm characteristics are Size, ROA, Volatility, Leverage, Loss Dummy,
Subordinated Debt Dummy, Interest Coverage, and Capital Intensity. The included
variables of governance attributes are Insider Ownership, Independent Director,
Institutional Ownership, Alpha, Earnings Quality, G—Index, and Delaware Dummy.
The deﬁnitions of all variables are provided in the Appendix. Since credit ratings
are ordinal, we conduct an Ordered Probit regression based on a seven categories of
ratings classiﬁcation to estimate the above model. We also include year and
industry dummies in the estimation of the model. The p-values are based on the
standard errors which are robust to heteroskedasticity and adjusted for ﬁrm
clustering.

Table 3.4 presents the results of our analysis about the effect of market
competition on credit ratings. Model 1 reports the results of an Ordered Probit
regression with the control variables of HHI, Return C orrelation and ﬁrm
characteristic, whereas Model 2 reports the results of an Ordered Probit regression
with all control variables including governance variables. The results from both
models Show that the estimated coefﬁcients on HHI index are positive and

statistically signiﬁcant at 5% and 10% level, respectively. The results support the

71

view that managers are more likely to be aligned with shareholders in more
competitive market and thus take riskier investments, which lowers credit ratings.
Model 2 Shows that the coefﬁcient estimate of Return Correlation is -0.163 and
statistically signiﬁcant at 1% level. To the extent that a correlation of a ﬁrm’s stock
returns with its industry counterparts reﬂects the interdependence of investment
opportunities, the ﬁnding indicates that predation risk lower credit rating.

For the explanatory variables of ﬁrm characteristics, Model 2 shows that the
estimated coefﬁcients on Size and Capital Intensity are Signiﬁcantly positive, while
the estimated coefﬁcients on Volatility and Subordinated Debt Dummy are
Signiﬁcantly negative. Model 2 also shows that the coefﬁcient estimates of ROA,
Leverage and Loss Dummy are 2.691, -1.530 and -0.415, respectively, and are
statistically signiﬁcant at 1% level. These results indicate that ﬁrms with higher
default risk have lower credit rating. However, the coefﬁcient estimate of Interest
Coverage has its expected sign but is not statistically signiﬁcant.

Regarding to the control variables of governance attributes, Model 2 reports.
that coefﬁcients on Insider Ownership and Alpha are negative and statistically
signiﬁcant at 5% and 10% level, respectively, which suggests that stock and stock
option holdings give managers risk-taking incentives and thus result in lower credit
rating. These ﬁndings are consistent with those of Ortiz-Molina (2008), who shows
positive association between borrowing costs and managerial stock and stock
option holdings. Consistent with the result with Bhojraj and Sengupta (2003), the
coefﬁcient estimate of Institutional Ownership is negative and statistically

Signiﬁcant at 1% level. This evidence is in line with the view that institutional

72

blockholders improperly inﬂuence managers for their private beneﬁts that are
harmful to minority Shareholders and bondholders or that institutional blockholers
inﬂuence managers for wealth transfer from bondholders to shareholders. We ﬁnd
that the estimated coefﬁcient on Delaware Dummy is negative and statistically
signiﬁcant. This result suggests that ﬁrms incorporated in takeover ﬁiendly states
like Delaware have lower credit rating because of a wealth transfer from
bondholders to shareholders in the event of takeover, which is consistent with
Francis, Hasan, Jonh and Waisman (2006). However, we do not ﬁnd a signiﬁcant
association between credit ratings and G-Index. Negative coefﬁcient on Earnings
Quality indicates that high quality of reported earnings contributes to the high
credit ratings. Finally, we do not ﬁnd a Signiﬁcant relation between Independent

Director and credit ratings.

3.6.2 Product Market Competition and Investment-Grade Credit Rating
To examine the effect of market competition on the likelihood of getting an
investment-grade credit rating, we use a Probit model. The dependent variable is
equal to one if credit rating is investment grade, and zero otherwise. Credit rating is
classiﬁed into investment grade if it is greater than or equal to BBB. The Probit
model include year and industry ﬁxed effects. The p-values are based on the
standard errors which are robust to heteroskedasticity and adjusted for ﬁrm
clustering.

Results are reported in Table 3.5. The estimated coefﬁcient on HHI index is

positive and statistically signiﬁcant at 5% level. This indicates that ﬁrms are more

73

likely to get investment-grade credit ratings in less competitive market since
managers who are not aligned with shareholders interests are more likely to avoid
risky projects in less competitive market. The coefﬁcient of Return Correlation is
negative and statistically signiﬁcant at 5% level, which suggests that predation risk
reduces the likelihood of getting an investment-grade credit rating. The coefﬁcient
estimate of Size and ROA are positive and statistically signiﬁcant, whereas the
coefﬁcient estimate of Volatility, Leverage, Loss Dummy, and Subordinated Debt
Dummy are negative and statistically Signiﬁcant. So far, the ﬁndings are similar to
the ﬁndings of the credit rating regression shown in Table 3.5. However, the
coefﬁcient estimate of Interest Coverage is statistically signiﬁcant at 10% level,
while it was insigniﬁcant in the credit rating regression shown in Table 3.4. Also
the estimated coefﬁcient on Capital Intensity becomes insigniﬁcant in this analysis
unlike the result in credit rating regression.

In terms of variables of governance attributes, the coefﬁcient on Insider
Ownership is negative but insigniﬁcant, whereas the coefﬁcient on Alpha is
negative and signiﬁcant. This suggests that stock option holdings by managers have
more negative effect on the likelihood of getting an investment-grade credit rating
than stock holdings by managers. The estimated coefﬁcient on Earnings Quality is
positive and statistically signiﬁcant, which indicates that high quality of reported
eamings increase the likelihood of getting an investment-grade credit rating.
Suprisingly, the estimated coefﬁcients on Institutional Ownership and Delaware
Dummy are not statistically different from zero. Those coefﬁcients were Signiﬁcant

in the credit rating regression in Table 3.4.

74

To explore the economic signiﬁcance of our ﬁndings, we compute the
change in probability of getting an investment-grade rating by moving from the 25th
to the 75'h percentiles of each control variable of interest while ﬁxing all other
variables at their means. If the variable of interest is a dummy variable, we compute
the change in probability of getting an investment-grade rating by changing the
value from 0 to 1 while ﬁxing all other variables at their means. The results are
presented in Table 3.7. Moving from the 25th to the 75th percentiles of HHI index
leads to increase in probability of getting an investment-grade rating by 0.031,
while moving from the 25th to the 75th percentiles of Return Correlation leads to
decrease in probability of getting an investment-grade rating by 0.039. Overall, the
results indicate that changes in ﬁrm characteristics lead to bigger changes in
probabilities of getting investment-grade ratings than changes in governance

attributes.

3.7 The Link between Product Market Competition and Risky Investments
We use the following speciﬁcation to investigate to what extent market competition
induces managers to conduct risk investments,
Risky Expenditures = )6” + B, HHI + ,8; Tobin 's Q + 6’3 Stock Return +
,64 Leverage +/I5 Size + [36 Surplus Cash +
,67 Sales Growth + ,8, Insider Ownership +
,b’g Independent Director + ,8”) Institutional Ownership +

1611 Equity Compensation + a

75

We include control variables based on the literature (e. g. Coles, Daniel and Naveen
(2006)). The independent variable is the sum of the expenditures in R&D and
advertising. If R&D or advertising expenditure is missing, we set it to zero. All
control variables are deﬁned in Appendix. Our coefficient of interest is ,6), which
measures to what extent market competition induces managers to conduct risk
investment. We expect a negative coefﬁcient on HHI since we conjecture that
managers are more likely to implement risky investments in more competitive
markets. We use two methodologies to estimate the above model. First, we use
Tobit for a ﬁrll sample of ﬁrms since Risky Expenditures have zero values in large
number of observations. Second, we use OLS for ﬁrms with nonzero Risky
Expenditures. The results are given in Table 3.7. Model 1 reports the results of
Tobit regression for a full sample of ﬁrms and Model 2 reports the results of OLS
regression for ﬁrms with nonzero Risky Investments. The results in Model 1 and
Model 2 Show that the coefﬁcients on HHI are negative and statistically signiﬁcant
at 5% level in both regressions. This results support the view that managers are
more likely to be aligned with shareholders in more competitive market and thus

take riskier investments.

3.8 Conclusion

In this paper, we investigate whether the link between product market competition
and managerial incentives affects ﬁrms’ credit ratings. Product market competition
can discipline self-interested managers as an effective govemance mechanism and

thus beneﬁt both shareholders and bondholders. However, agency theory by Jensen

76

and Meckling (1976) suggests that managers closely aligned with shareholders may
take actions that can transfer wealth from creditors to Shareholders.

We ﬁnd that market competition has positive relation to shareholder value
but negative relation to credit ratings. We also ﬁnd that increase in market
competition leads to decrease in the likelihood to receive investment-grade credit
ratings. Finally, we ﬁnd that ﬁrms in more competitive market tend to make more
R&D and advertising expenditures. Overall, our results indicate that managers are
more likely to be aligned with shareholders in more competitive markets and thus
undertake riskier investments to maximize shareholder value, which lowers ﬁrms’
credit ratings. Our ﬁndings suggest that credit agencies rationally assign ﬁrm’s
credit rating considering the ﬁrm’s future choice of risk level that is reﬂected in the
level of product market competition. Our study contributes to the literature by
providing evidence that that product market competition align more closely the
interests of managers and shareholders and thus create managerial incentives to
undertake riskier investments to maximize shareholder value, which lowers ﬁrms’

credit ratings.

77

Chapter 3 Appendix

78

A3.1 Variable Deﬁnitions

V ariablcs
Credit Rating

Investment Grade

Adjusted Tobin’s Q

Risky Expenditures

Size

ROA

Volatility

Leverage

Loss Dummy

Subordinated Debt
Dummy

Interest Coverage

Capital Intensity
Firm Age
S&P 500 Index Dummy

Insider Ownership

Independent Director

Institutional Ownership

Deﬁnitions

S&P credit ratings (data280) are converted into credit rating scores as follows. It
is 1 for rating <= CCC", 2 for B'<= rating <= Bi, 3 for BB' <= rating <= 83’, 4
333' <=rating <= BBB+, 5 for A' <= rating <= AI, 6 for AA’ <= rating <= AAI,
and 7 for rating = AAA.

1 if the ﬁrm’s credit rating is greater than or equal to BBB, 0 otherwise.

Industry adjusted Tobin's Q is the difference between the ﬁrm‘s Q and the
median Q in the same industry as deﬁned by three-digit SIC code. Q is deﬁned
by the market value of assets divided by book value of assets and calculated as
(data6+data l 99*data25-data60)/data6.

The sum of the expenditures in R&D (data46) and advertising (data45) divided
by total assets (data6). If R&D or advertising expenditure is missing, we set it to
zero.

Natural logarithm of total assets (data6) (in millions of dollars).

Net income before extraordinary items divided by total assets calculated as
data 1 8/data6.

Standard deviation of monthly stock returns over the past 5 years.

The sum ofdebt in current liabilities plus long term debt divided by total assets
calculated as (data34+data9)/data6.

1 if ROA is negative in the current and prior ﬁscal year, 0 otherwise.

1 ifthe ﬁrm has subordinated debt (data80). 0 otherwise.

Operating income before depreciation divided by interest expense calculated as
data 13/(data15 or data339).

Net PPE divided by total assets calculated as data8/data6.

The number of years since the ﬁrm has ﬁrst appeared on CRSP.

1 ifthe firm is included in S&P 500 index, 0 otherwise.

The aggregate percentage ownership of common stocks held by top ﬁve
executives (ofﬁcers and directors).

A percentage ofthe independent directors on the board.

The aggregate percentage ownership of shares held by institutional investors
which own at least 5% Of outstanding shares (in millions of shares).

79

A3.1 Variable Deﬁnitions (Continue)

Alpha

Earnings Quality

G-Index
Delaware Dummy

Equity Compensation

HHI

Stock Return

Suplus Cash

Sales Growth

Return Correlation

The sum of the number of common stock, restricted stock and unexercised stock
options held by top ﬁve executives (ofﬁcers and directors) as a percentage of
total shares outstanding.

Negative one times the loadings on the accrual quality (AQ) factor by estimating
following Fama-French three-factor model augmented with AQ factor of Ecker,
Francis, Kim, Olsson, and Schipper (2006):

R, = a + ,BIMKT + BzSMB + B3HML + 1mg

R, is ﬁrm‘s daily excess stock returns. AQ is calculated as the standard
deviation of the residuals by estimating the regressions of the total current
accruals on past, current, and ﬁiture cash ﬂow from operations, change in
revenues, and gross value of property, plant, and equipment.

Shareholder rights score of Gompers, Ishii, and Metrick (2003)
1 if the ﬁrm is incorporated in Delaware, 0 otherwise

The sum of the Black-Scholes value of new stock options received by the top
ﬁve executives (ofﬁcers and directors) as a percentage of total compensation
paid to them. Total compensation is comprised of salary, bonus, other annual
compensation, value of restricted stock granted, Black-Scholdes value of new
stock options granted, long-term incentive payouts, and all other compensations.

Hirﬁndahl-Hirschman index is computed as the sum of squared market shares,

'VI 2
HHIJ, — i=1 sq,

where i, j, and t denote ﬁrm, industry, and year, respectively. Market shares are
based on ﬁrms’ sales (data12). Industry is classified b} three-digit SIC codes.

Cumulative daily stock returns over the ﬁscal year.
Available cash amount to ﬁnance new projects divided by total assets calculated
as (data308-data125+data46)/data6.

Natural logarithm ofthe ratio of sales (data12) in the current ﬁscal year to sales
in the prior ﬁscal year.

The proxy for the correlation of ﬁrm stock returns with its corresponding

industry stock returns is the estimated coefﬁcient on the industry return index
from the following regression of Haushalter, Klasa. and Maxwell,

rt : B!) + ﬁll}! + )8} 1nd

where r, is a iirm’s monthly stock returns for our sample period. Ii, is the
monthly equally-weighted market return, In“, is the monthly equally-weighted
index return for the same industry as deﬁned by two-digit SIC code.

80

Table 3.1
Credit Rating Categories

Firm credit ratings are obtained from S&P’s long-terrn issuer credit ratings in Compustat (data280).
S&P’s ratings are based on its assessment of the creditworthiness of a ﬁrm as a senior debt obligor.
Rating is classiﬁed into speculative grade if it is below 888-, and investment grade otherwise.

 

 

Converted Credit
Rating Score S&P Credit Rating Grade
7 AAA Investment
6 AA+, AA, AA- Investment
5 A+, A, A- Investment
4 BBB+, BBB, BBB- Investment
3 88+, BB, BB- Speculative
2 B+, B, B- Speculative
1 CCC+, CCC. CC. C, D, SD Spgculative

 

81

Table 3.2
Summary Statistics

This table shows the summary statistics for the variables used in the analysis. The sample consists of
6.957 ﬁrm year observations from 1996 to 2006. Variable deﬁnitions are in the Appendix.

 

 

Variables N Mean Std. Q1 Median Q3
Independent Variables
Credit Rating 6589 4.08 1.07 3 4 5
Investment Grade 6589 0.72 0.45 0 1 1
Adjusted Tobin's
Q 6589 0.25 1.12 -0. 14 0.03 0.38
Firm Characteristics
Total Assets 6589 17140 66633 1657.21 3774.66 10947
ROA 6589 0.04 0.09 0.01 0.04 0.07
Volatility 6589 0.11 0.05 0.07 0.10 0.13
Leverage 6589 0.31 0.17 0.19 0.30 0.39
Loss Dummy 6589 0.06 0.24 0 0 0
Subordinated Debt
Dummy 6589 0.16 0.36 0 0 0
Interest Coverage 6589 15.32 99.24 3.72 6.37 12.13
Capital Intensity 6589 0.34 0.24 0.15 0.29 0.53
Firm Age 6589 31.58 21.56 13 29 43
S&P 500 Index
Dummy 6589 0.45 0.50 0 0 1
Governance Attributes
Insider Ownership 6589 1.69 5.11 0 0 0.71
Independent
Director 6589 67.66 17.35 57.14 70 81.25
Institutional
Ownership 6589 8.45 9.01 0 6.69 13.10
Alpha 6589 2.22 2.96 0.80 1.63 2.90
Eamings Quality 6589 0.07 0.32 -0.07 0.10 0.25
G-lndex 6589 9.76 2.60 8 10 12
Delaware Dummy 6589 0.57 0.49 0 1 1
Equity
Compensation 6575 34.02 23.98 14.78 32.42 51.19
Market Competition
HHI 6589 0.16 0.15 0.06 0.11 0.21
Return Correlation 6589 0.82 0.80 0.33 0.91 1.31

 

82

Table 3.3
Regression analysis of the effect of market competition on Tobin's Q

This table shows OLS regression results of the effects of market competition on the Tobin’s Q. The
dependent variable is the adjusted Tobin’s Q. All variable deﬁnitions are in the Appendix. The
coefﬁcients on the intercepts are not reported. All models include industry dummies (based on 48
industry classiﬁcation by Fama and French) and year dummies, whose coefﬁcient estimates are not
reported. P-values in parenthesis are corrected for heteroscedasticity and ﬁrm clustering.

 

 

Variables Model 1 Model 2
Market Competition
HHI -0.424 -0.384
(0.008) (0.016)
Firm Characteristics
Size -0.069 -0.097
(0.007) (0.000)
ROA 3.183 3.190
(0.000) (0.000)
Leverage 0.108 0.159
(0.720) (0.584)
Log (ﬁrm age) -0.076 -0.040
(0.037) (0.244)
S&P 500 Index Dummy 0.448 0.424
(0.000) (0.000)
Govemance Attributes
Insider Ownership 0.001
(0.797)
Independent Director 0.002
(0.270)
Institutional Ownership -0.008
(0.001)
Equity Compensation 0.006
(0.000)
G-lndex -0.023
(0.015)
Delaware Dummy 0.071
(0.181)
Year Dummies Yes Yes
Industry Dummies Yes Yes
R-Squared 0.193 0.214
Number of Observations 65 89 6575

 

83

Table 3.4
Regression analysis of the effect of market competition on credit ratings

This table shows Ordered Probit regression results of the effects of market competition on the credit
rating. The dependent variable is the credit rating. All variable deﬁnitions are in the Appendix. The
coefﬁcients on the intercepts are not reported. All models include industry dummies (based on based on
48 industry classiﬁcation by Fama and French) and year dummies, whose coefﬁcient estimates are not
reported. P-values in parenthesis are corrected for heteroscedasticity and ﬁrm clustering.

 

 

Predicted
Variables Sign Model 1 Model 2
Market Competition
HHI + 0.530 0.490
(0.043) (0.054)
Predation Risk
Retum Correlation - -0.161 -0.163
(0.000) (0.000)
F inn Characteristics
Size + 0.472 0.453
(0.000) (0.000)
ROA + 2.887 2.691
(0.000) (0.000)
Volatility - -l7.050 -16.587
(0.000) (0.000)
Leverage - -1.584 -1.530
(0.000) (0.000)
Loss Dummy - -0.427 -0.415
(0.000) (0.000)
Subordinated Debt Dummy - -0.493 -0.451
(0.000) (0.000)
Interest Coverage + 0.000 0.000
(0.346) (0.341)
Capital Intensity + 0.545 0.496
(0.012) (0.023)
Govemance Attributes
Insider Ownership - -0.014
(0027)
Independent Director + 0.000
(0.950)
Institutional Ownership ? -0.010
(0.000)
Alpha - -0.043
(0.068)
Earnings Quality + 0.238
(0.001)
G—Indcx '? -0.020
(0.123)
Delaw are Dummy - -0. l 30
(0059)
Year and Industry Dummies Yes Yes
Pseudo R-Squared 0.329 0.339
Number of Observations 65 89 65 89

 

84

Table 3.5
The effect of market competition on receiving an investment grade credit rating

This table shows Probit regression results of the effects of market competition on receiving investment grade
rating. The dependent variable is investment grade rating. All variable deﬁnitions are in the Appendix. The
coefﬁcients on the intercepts are not reported. All models include industry dummies (based on 48 industry
classiﬁcation by Fama and French) and year dummies, whose coefﬁcient estimates are not reported. P-values in

parenthesis are corrected for heteroscedasticity and ﬁrm clustering.

 

 

Variables Predicted Sign Estimate
Market Competition
HHI + 0.838
(0.022)
Predation Risk —
Return Correlation -0. 161
(0.029)
Firm Characteristics
Size + 0.532
(0.000)
ROA + 5.134
(0.000)
Volatility - -20.239
(0.000)
Leverage - -2.854
(0.000)
Loss Dummy - -0.718
(0.000)
Subordinated Debt Dummy - -0.807
(0.000)
Interest Coverage + 0.000
(0.085)
Capital Intensity + 0.251
(0.470)
Governance Attributes
Insider Ownership - -0.009
(0.390)
Independent Director + -0.001
(0.694)
Institutional Ownership ? -0.001
(0.879)
Alpha - -0.062
(0.001 )
Earnings Quality + 0.209
(0.046)
G—Index ? 0.007
- (0.728)
Delaware Dummy - -0.121
((1.242)
Year & Year Dummies Yes
Pseudo R-Squared 0.551
Number of Observations 6552

 

85

Table 3.6
Probability changes in receiving an investment grade credit ratings

This table shows probability change of receiving an investment grade rating when the variable of
interest is changed from 25‘h to 75‘h percentiles, ﬁxing all other variables at their mean values. If
the variable of interest is dummy variable, it is changed from 0 to 1. Variable deﬁnitions are in the
Appendix.

 

 

Variables Predicted Sign Probability Change
Market Competition

HHI + 0.031
Predation Risk

Return Correlation - -0.039

Firm Characteristics

Size + 0.234
ROA + 0.069
Volatility - -O.238
Leverage - -0.136
Loss Dummy - -0.227
Subordinated Debt Dummy - -0.248
Interest Coverage + 0.001
Capital Intensity + 0.024

Govemance Attributes

Insider Ownership - -0.002
Independent Director + —0.006
Institutional Ownership ? -0.002
Alpha - 0031
Earnings Quality + 0.017
G-Index ‘? 0.007
Delaware Dummy - -0.030

 

86

Table 3.7
Regression analysis of the effect of market competition on expenditures
in R&D and Advertising

This table shows regression results of the effects of market competition on the sum of the R&D and
advertising expenditures. The dependent variable is the sum of the expenditures in R&D (data46) and
advertising (data45) divided by total assets (data6). If R&D or advertising expenditure is missing, we set it
to zero. Model 1 is a Tobit regression. Model 2 is a OLS regression with nonzero sum of expenditures in
R&D and advertising. All variable deﬁnitions are in the Appendix. The coefﬁcients on the intercepts are
not reported. All models include industry dummies (based on 48 industry classiﬁcation by Fama and
French) and year dummies, whose coefﬁcient estimates are not reported. P-values in parenthesis are
corrected for heteroscedasticity in Model 1 and for heteroscedasticity and ﬁrm clustering in Model 2.

 

 

Variables Model 1 Model 2
Market Competition
HHI -0.0259 -0.0268
(0.000) (0.015)
Firm Characteristics
Tobin's Q 0.0058 0.0052
(0.000) (0.009)
Stock Return -0.0084 -0.0081
(0.000) (0.000)
Leverage 0.0252 0.0252
(0.000) (0.207)
Size -0.0029 -0.0028
(0.000) (0.047)
Surplus Cash 0.1657 0.1654
(0.000) (0.000)
Sales Growth -0.0227 -0.0230
(0.000) (0.002)
Govemance Attributes
Insider Ownership 0.0002 0.0005
(0.203) (0.329)
Independent Director 0.0002 0.0002
(0.000) (0.139)
Institutional Ownership -0.0002 -0.0002
(0.049) (0.195)
Equity Compensation 0.0002 0.0003
(0.000) (0.000)
Year Dummies Yes Yes
Industry Dummies Yes Yes
Number of Observations 3370 2938

 

87

BIBLIOGRAPHY

88

Alchian, Annen A., 1950, Uncertainty, evolution, and economic theory, Journal of
Political Economy 58, 211-221.

Allen, Franklin, and Douglas Gale, 1999, Corporate governance and competition,
Working Paper.

Ashbaugh-Skaife, Hollis, Daniel W. Collins, and Ryan LaFond, 2006, The effects of
corporate governance on ﬁrms’ credit ratings, Journal of Accounting and Economics 42,
203-243.

Bhagat, Sanjai, and Bernard S. Black, 2000, Board independence and long-term
performance, Working Paper.

Bhojraj, Sanjeev, and Partha Sengupta, 2003, Effect of corporate governance on bond
ratings and yields: The role of institutional investors and outside directors, Journal of
Business 76, 455-475.

Bradley, Michael, Dong Chen, George Dallas, and Elizabeth Snyderwine, 2007, The
relationship between corporate governance and credit risk, bond yields and ﬁrm valuation,
Working Paper.

Barber, B. M., and J. D. Lyon, 1996, Detecting abnormal operating performance: The
empirical power and speciﬁcation of test-statistics, Journal of Financial Economics
41, 359-399.

Bates, Thomas W., 2005, Asset sales, investment Opportunities, and the use of
proceeds, Journal of Finance 60, 105-135.

Bebchuk, Lucian A., and Alma Cohen, 2005, The cost of entrenched boards, Journal
of Financial Economics 78, 409—433.

Carhart, Mark, 1997, On Persistence in mutual fund perfomiance, Journal of Finance 52,
57-82.

Coles, Jeffrey L., Naveen D. Daniel, and Lalitha Naveen, 2006, Managerial incentives
and risk-taking, Journal of Financial Economics 79, 431-468.

Danes, Robert, 2001, Does Delaware law improve ﬁrm value‘.’, Journal of Financial
Economics 62, 525-558.

Eckbo. B. Espen, and Ronald W. Masulis, and Oyvind Norli, 2000, Seasoned public
Offerings: Resolution of the ‘new issues puzzle’, Journal ofFinancial Economics 56,

251-291.

Ecker, Frank, Jennifer Francis, Irene Kim, Per M. Olsson, and Katherine Schipper, 2006,
A return-based representation Of earnings quality, The Accounting Review 81, 749-780.

89

Faleye, Olubunmi, 2007, Classiﬁed boards, Firm Value, and Managerial Entrenchment,
Journal of Financial Economics 83, 501-529.

Francis, Bill B., Iftekhar Hasan, Kose John, and Maya Waisman, 2006, Does Delaware
incorporation affect the ﬁrm’s cost of debt capital?, Working Paper.

Friedman, Milton, 1953, The methodology of positive economics, published in Essays on
Positive Economics, University of Chicago Press.

Ferreira, Miguel A, and Paul A, Laux, 2007, Corporate boards and SEOS: The effect
of certiﬁcation and monitoring, Working Paper.

Fu, Fangjian, Overinvestment and the operating performance of SEO ﬁrms, Working
Paper.

Gompers, Paul A., Joy Ishii, and Andrew Metrick, 2006, Extreme governance: An analysis of
dual-class ﬁrms in the United States, Working Paper.

Giroud, Xavier, and Holger M. Mueller, 2008, Corporate governance, Product Market
Competition, and Equity Prices, Working Paper.

Gompers, Paul A., Joy L. Ishii, and Andrew Metrick, 2003, Corporate govemance and
equity prices, Quarterly Journal of Economics 118, 107-155.

Grifﬁth, Rachel, 2001, Product market competition, efﬁciency and agency cost: An
empirical analysis, Working Paper.

Grullon, Gustavo, and Roni Michaely, 2007, Corporate payout policy and product market
competition, 2007, Working Paper.

Haushalter, David, Sandy Klasa, and William F. Maxwell, 2009, The inﬂuence of
product market dynamics on the ﬁrm’s cash holdings and hedging behavior, Journal of
Financial Economics, Forthcoming.

Jensen, Michael C ., and William H. Meckling, 1976, Theory of the ﬁrm: Managerial
behavior, agency costs and ownership structure, Journal of Financial Economics 3,
305-360.

Jensen, Michael C ., 1986, Agency costs of free cash ﬂow, corporate ﬁnance and
takeovers, American Economic Review 76, 323-329.

Jung, Kooyul, Yong Cheol Kim, and René M. Stulz, 1996, Timing, investment
opportunities, managerial discretion, and the security issue decision, Journal of
Financial Economics 42, 159-185.

Klock, Mark S.. Sattar A. Mansi, and William F. Maxwell, 2005, Does corporate
governance matter tO bondholders?, Journal of Financial and Quantitative Analysis 40,
693-719.

90

Kim, E. Han, and Amiyatosh Pumanandam, 2006, Why do investors react negatively
to seasoned equity offerings?, Working Paper, University of Michigan.

Koppes, Richard R, Lyle G. Ganske, and Charles T. Haag, 1999, Corporate
governance out of focus: The debate over classiﬁed boards, Business Lawyer 54,
1023-1044.

Loughran, Tim, and Jay Ritter, 1997, The Operating performance of ﬁrms conducting
seasoned equity offerings, Journal of Finance 52, 1823-1850.

Lee, Inmoo, 1997, Do ﬁrms knowingly sell overvalued equity?, Journal of Finance
32, 1439-1466.

Leland, Hayne E., and David H. Pyle, 1977, Informational asymmetries, ﬁnancial
structure, and ﬁnancial interrnediation, Journal of Finance 32, 371-387.

Loughran, Tim, and Jay Ritter, 1995, The new issue puzzle, Journal of Finance 50,
23-51.

Masulis, Ronald W., Cong Wang, and Fei Xie, 2007, Corporate governance and
acquirer returns, Journal of Finance 62, 1851-1889.

Myers, Stewart C ., and Nicholas S. Majluf, 1984, Corporate ﬁnancing and investment
decisions when ﬁrms have information that investors do not have, Journal of
Financial Economics 13, 187- 221.

Megginson, William L., Scott B Smart, and Chad J. Zutter, 2008, The evolution of equity
ﬁnancing: A comparison of dual-class and single-class SEOS, Working Paper.

McConnell, John J., and Henri Servaes, 1990, Additional evidence on equity ownership
and corporate value, Journal of Financial Economics 27, 595-612.

Morck, Randall, and Fan Yang, 2001, The mysterious growing value of the S&P 500
Membership, Working Paper.

Ortiz-Molina, Heman, 2006, Top-management incentives and the pricing of corporate
public debt, Journal of Financial and Quantitative Analysis 41 , 317-340.

Richardson, Scott, 2006, Over-investment of free cash ﬂow, Review of Accounting
Studies 11, 159-189.

Smith, James F., 1977, The equal credit opportunity act of 1974: A cost/beneﬁt
analysis, Journal ofFinance 32, 609-622.

Stulz, Rene, I990, Managerial discretion and optimal ﬁnancing policies, Journal of
Financial Economic, 26, 3-27.

91

Shleifer, Andrei, and Robert Vishny, 1997, A survey of corporate governance, Journal of
Finance 52, 737-783.

Subramanian, Guhan, 2004, The disappearing Delaware effect, Journal of Law,
Economics & Organization 20, 32-59.

Xiumin Martin, 2007, Product market competition and the valuation of ﬁrm cash
holdings, Working Paper.

92

 

Wlilljlljljljljljljjljllljljjllf