Eu...

.3? ELF 35. ..

ammﬁmw.
H.u
. . .

z .210!
9.3.. E!
i x ,1

 

 

3.4.3.

 

 

era

This is to certify that the
dissertation entitled

Three Essays on Credit Contracts

presented by

Sung-guan Yun

has been accepted towards fulﬁllment
of the requirements for the

PhD. degree in Economics

 

 

If N INK/{UH 4/14/71;

 

Major Professor’s Signature
4% 2 j /2 o 4 51

Date

MSU is an Afﬁrmative Action/Equal Opportunity Employer

 

—.-d-.-.-.-.‘

 

‘_—.— -.-.-‘-.-.--_-<---A-i—

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

 

. 2
MARglsg 2m

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5108 IUProj/AccaPresIClRCIDateDuejndd

 

 

 

 

 

THREE ESSAYS ON CREDIT CONTRACTS
By

Sung-guan Yun

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Economics

2010

 

 

 

 

4 _.__ r—_.~._—

ABSTRACT

THREE ESSAYS ON CREDIT CONTRACTS

By

Sung-guan Yun

This study contains three essays. In the ﬁrst chapter, we tested the claim that in
emerging economies institutional and regulatory constraints deter lenders from moni-
toring and disciplining ﬁrms by studying the experience of South Korean chaebols and
their reform of the late nineties. Building on Holmstrom and Tirole (1997) and Suﬁ
(2007), we argue that when a ﬁrm is granted a syndicated loan the lead bank retains a
large share of the loan to commit to monitoring the borrower. We ﬁnd evidence that
the institutional arrangements of chaebols - especially the bailout protection offered
by the South Korean government - diluted lenders’ incentives to monitor chaebol
ﬁrms, as reflected by the concentration of syndicated loans. However, after the re-
form dismantled the chaebol safety net in the late nineties, banks stepped up their
monitoring effort (and toughened lending standards). Although internal governance
mechanisms tend to compensate for the lack of creditors’ monitoring, in our data
such substitutability appears to be only partial.

In the second chapter, I investigated the effects of the protection of creditors’ rights
on the structure of syndicated loans, such as the shares of the loan held by the lead
arrangers and the loan distribution between the lead arrangers and the participant

lenders. I found that the average loan concentration (proxy for the intensity of mon-

itoring efforts by the lead arrangers) of the ﬁrm, which sets up more provisions for
protecting creditor’s rights, is lower and the likelihood of the creditors’ holding col—
lateral is smaller. However the extent of the impact of such protection on the loan
concentration also depends on the degree of the product market competition the ﬁrm
faces. For the ﬁrms that belong to the highly competitive product markets, I found
that the impact of the G index is pronounced and the creditors are less likely to hold
collateral, and embed fewer ﬁnancial covenants when the ﬁrms are prone to protect
creditors’ rights.

In the last chapter, I examine the impact of ﬁrms’ asset liquidation value which
is proxied by asset liquidity on debt contracting using comprehensive US syndicated
loan contracts. I employ various measures of the asset liquidity at the three-digit
industry level. The borrower, belonging to a higher asset liquidity industry at the
time of loan origination, experiences a lower loan spread. However as the maturity of
a loan increases the impact of the asset liquidity declines. Furthermore I found that
the loan spreads are affected by the number of lenders because of the concern about
coordination failures, and by the possibility of the borrower’s being involved in M&A
activities. These results are consistent with incomplete contracting and transaction

cost theories of liquidation value and ﬁnancial structure.

To my wife, two daughters, my parents, father-in—law, and

mother-in-law for their love and encouragement

 

ACKNOWLEDGMENTS

First, I would like to thank my adviser, Dr. Raoul Minetti, for his encourage-
ment, suggestion, and guidance. Throughout my doctoral work, I learned invaluable
lessons from him by working with him and learning in his classes. The numerous
conversations in his office and coffee shops will not be forgotten.

I also want to thank Dr. Luis Araujo for his encouragement and teaching. What
I learned in his class will remain as my valuable assets, and I enjoyed his lectures. I
thank Dr. Susan Chun Zhu and Dr. Kirt C. Butler. They generously accepted to be
the committee member. Especially, their encouraging words and ideas helped me to
shape my dissertation.

I also thank Dr. Duncan Boughton for his encouragement and invaluable help in
starting my ﬁrst chapter. I am very much grateful to my church members in University
Reformed Church for their love, prayer, and forgiveness in Christ Jesus.

I would like to thank my colleagues in the PhD program for all of their help and
support. Especially the conversations with Young-J00 Choi were very enjoyable and
stimulating in developing ideas.

Last but not least, I would also like to thank my family, including my wife, Hye-
sun Choe, my children, Hajin and Jinsun, my parents, father-in—law, and mother-in-
law, for their enduring patience and support as I completed this dissertation. This

dissertation is dedicated to them.

TABLE OF CONTENTS

List of Tables .......................................................................................................... viii
List of Figures ........................................................................................................... ix
1 Institutions, Bank Monitoring, and the Structure of Credit Con-
tracts: Evidence from Korean Chaebols ............................................................................ 1
1.1 Introduction ........................................................................................................................... 2
1.2 Prior Literature ......................................................................................................................... 6
1.3 Institutional Background ....................................................................................................... 7
1.4 Theoretical Hypotheses ........................................................................................................... 11
1.4.1 Baseline Hypotheses ................... . ....................................................................................... 11
1.4.2 Other Hypotheses .................................................................................................................. 13
1.5 Data and Measurement ........................................................................................................... 14
1.5.1 Preliminary Observations ................................................................................................. 15
1.5.2 Data Description ................................................................................................................. 15
1.5.3 Measurement .......................................................................................................................... 17
1.5.4 Sample Properties ................................................................................................................. 20
1.6 Empirical Tests and Results ............................................................................................... 22
1.6.1 Chaebols and Syndicate Structure ................................................................................. 22
1.6.2 Chaebols and Foreign Banks ........................................................................................... 25
1.6.3 The Role of the Safety Net .............................................................................................. 27
1.6.4 Internal and External Governance ................................................................................. 30
1.6.5 Monitoring or Diversiﬁcation? ........................................................................................ 32
1.7 Further Tests ........................................................................................................................... 33
1.7.1 Robustness ............................................................................................................................ 33
1.7.2 Other Loan Characteristics .............................................................................................. 35
1.8 Conclusion ................................................................................................................................ 36
1.9 Bibliography ............................................................................................................................. 83
2 Creditors’ Rights and Loan Structure ............................................................................ 87
2.1 Introduction ............................................................................................................................. 88
2.2 Background .............................................................................................................................. 93
2.2.1 Protection of Creditors’ Rights and Structure of Loans ......................................... 93
2.2.2 Product Market Competition and Managerial Incentive ........................................ 96
2.3 Data and Statistics ................................................................................................................ 97
2.4 Empirical Speciﬁcation and Key Variables ..................................................................... 101
2.5 Results ..................................................................................................................................... 102
2.5.1 Preliminary results ........................................................................................................... 102
2.5.2 Risk-shifting and Loan structure .................................................................................. 105
2.5.3 Asset Substitutions and Loan structure ....................................................................... 107
2.5.4 CEO’s Compensation Structure ...................................................................................... 109
2.5.5 Impact of State’s Laws on Loan Contracts ............................................................... 110
2.5.6 Product Market Competition ........................................................................................ 111
2.6 6 Endogeneity ........................................................................................................................ 113

vi

 

 

2.7 Conclusion .............................................................................................................................. 113

2.8 Bibliography ........................................................................................................................... 128
3 The Effects of Asset Liquidity on the Loan Pricing .................................................... 131
3.1 Introduction ........................................................................................................................... 132
3.2 Measures of Asset Liquidity .............................................................................................. 133
3.3 Sample Construction, Empirical Speciﬁcation ............................................................. 134
3.3.1 Sample Construction ........................................................................................................ 134
3.3.2 Empirical Speciﬁcation and Key Variables ............................................................... 134
3.3.3 Summary Statistics ........................................................................................................... 136
3.4 Empirical Results ................................................................................................................. 138
3.4.1 Main results ........................................................................................................................ 138
3.4.2 Maturity and Uncertainty ................................................................................................ 140
3.4.3 Coordination failure and Pricing .................................................................................... 140
3.5 Discussion and Robust test ............................................................................................... 141
3.6 Conclusion .............................................................................................................................. 143
3.7 Bibliography ........................................................................................................................... 156

vii

 

LIST OF TABLES

1.1 Summary Statistics for Syndicated Loan Deals ............................................... 39
1.2 Summary Statistics for Syndicated Loan Deals by Chaebols ....................... 41
1.3 Chaebols and Syndicated Loan Structure .......................................................... 47
1.4 Chaebols and Syndicated Loan Structure .......................................................... 53
1.5 Chaebols and Foreign Lenders .............................................................................. 55
1.6 Domestic Bank Equity Participation .................................................................. 57
1.7 Intra—Chaebol Effects ................................................................................................. 61
1.8 Internal and External Governance ......................................................................... 63
1.9 Internal and External Governance in Chaebols ............................................... 67
1.10 {Internal and External Governance in Chaebols ............................................. 71
1.11 Robustness Tests ....................................................................................................... 75
1.12 Loan Characteristics and Chaebols ...................................................................... 78
1.13 Appendix : Variables and Sources ........................................................................ 80
2.1 Descriptive Statistics ................................................................................................ 115
2.2 Creditor’s Rights and Loan Structure ................................................................. 117
2.3 Risk-Shifting and Loan Structure ......................................................................... 119
2.4 Asset Substitutions and Loan Structure ............................................................ 121
2.5 CEO compensation and Loan Structure ............................................................. 122
2.6 State Law and Loan Structure .............................................................................. 123
2.7 Market Competition and Loan Structure .......................................................... 124
2.8 Market Competition and Loan Structure (Cont.) ......................................... 126
3.1 Summary Statistics ................................................................................................... 145
3.2 Effects of Asset Liquidity on the Loan Pricing .............................................. 146
3.3 Default Probability and Asset Liquidity ............................................................ 147
3.4 Maturity and the Loan Pricing ............................................................................. 149
3.5 Coordination Failure and the Loan Pricing ...................................................... 151
3.6 Leverage Increasing Activity and the Loan Pricing ..................................... 153
3.7 Takevover and the Loan Pricing ........................................................................ 155

viii

LIST OF FIGURES

1 Volume of Syndicated Loans in Korea and Asia—Paciﬁc Countries ............. 38

2 Relationship with the G index and the loan concentration measures ...... 115

 

CHAPTER 1

Institutions, Bank Monitoring, and
the Structure of Credit Contracts:

Evidence from Korean Chaebols

Abstract

It is often claimed that in emerging economies institutional and regulatory con-
straints deter lenders from monitoring and disciplining ﬁrms. This paper tests this
claim studying the experience of South Korean chaebols and their reform of the late
nineties. Building on Holmstrom and Tirole (1997) and Suﬁ (2007), we argue that
when a ﬁrm is granted a syndicated loan the lead bank retains a large share of the
loan to commit to monitoring the borrower. We ﬁnd evidence that the institutional
arrangements of chaebols - especially the bailout protection offered by the South Ko-
rean government - diluted lenders’ incentives to monitor chaebol ﬁrms, as reflected
by the concentration of syndicated loans. However, after the reform dismantled the
chaebol safety net in the late nineties, banks stepped up their monitoring effort (and
toughened lending standards). Although internal governance mechanisms tend to

compensate for the lack of creditors’ monitoring, in our data such substitutability
1

appears to be only partial.

1. 1 Introduction

The institutional environment of a country and the performance of its credit market
are tied to each other. The quality of laws and policies shapes incentives and con-
straints of ﬁrms and lenders and, hence, their behavior in credit transactions. One
prominent feature of the credit market that institutions allegedly affect is lenders’
monitoring of borrowers. This monitoring activity is a fundamental role of ﬁnancial
intermediaries: by disciplining borrowers, ﬁnancial intermediaries can channel funds
towards efficient projects and mitigate ﬁrms’ misbehavior (Diamond, 1984; Holm-
strom and Tirole, 1997; Carletti, 2004).

The impact that poor institutions have on the discipline exerted by creditors is
perceived as an acute problem in emerging economies. For example, to ease the
growth of the corporate sector, the governments of emerging economies frequently
adopt bailout policies that protect large corporations or business conglomerates from
failure. This safety net is blamed for diluting lenders’ incentives to monitor ﬁrms (see,
e.g., Vives, 2006). In emerging economies, lenders’. monitoring can also be inhibited
by lack of good quality information, for example due to poor accounting rules. In
turn, lenders’ inadequate monitoring can nurture ﬁrms’ inefficiency, such as excessive
risk taking. Indeed, scholars and policy-makers maintain that a cost of bailout policies
is that they weaken market discipline and distort credit allocation (OECD, 2000).1

The objective of this paper is to shed new light on the impact of institutions on the
monitoring role of the credit market in emerging economies. We study the experience

of South Korean business conglomerates (chaebols), an ideal testing ground for our

 

1A broad body of literature investigates the role of ﬁnancial market development for resource
allocation and economic growth (see, e.g., Rajan and Zingales, 1998; Beck, Levine, and Loayza,
2000; Wurgler, 2000; Allen and Gale, 2000). Demirgiic-Kunt and Maksimovic (1998) and Levine
(1999) ﬁnd that strong investor protection and legal enforcement favor economic growth.

2

purposes. Before 1997, the institutional environment in which chaebol ﬁrms oper-
ated sharply differed from that of non-chaebol ones. Distressed chaebol ﬁrms were
protected by a safety net consisting of an implicit government bailout policy and
a system of cross-debt payment guarantees among chaebol afﬁliates. Furthermore,
chaebol ﬁrms had lower accounting transparency than non-chaebol ones. These insti-
tutional arrangements allegedly discouraged lenders from monitoring and disciplining
chaebol ﬁrms (Lim, Haggard, and Kim, 2003; Nam, Kang, and Kim, 1999). Following
the economic crisis that hit South Korea between the end 1997 and the beginning
of 1998, this institutional environment was subject to a policy shock, the reform en-
acted by the South Korean government under the external pressure of the IMF and
the World Bank. The government prohibited cross-debt guarantees among chaebol
affiliates and abandoned its bailout policy by letting some chaebols go bankrupt.
Moreover, it adopted drastic measures to step up the transparency of chaebols, thus
making available to lenders more instruments to monitor chaebol ﬁrms.

We investigate the impact that the institutional arrangements of chaebols and their
reform had on banks’ monitoring. There is a consensus in the literature that creditors’
monitoring incentives can be inferred from the design of credit contracts (see, e.g.,
Suﬁ, 2007). Building on this premise, we match detailed information on chaebols and
their reform with rich contract-level data from the Korean syndicated loan market,
as well as borrower speciﬁc information. The syndicated loan market is ideally suited
to study lenders’ monitoring activity. A syndicated loan is provided by a group of
banks to a borrower. In a syndication, a borrower designates a lead arranger(s), who
then turns to potential participant lenders for a co-ﬁnancing of the loan; in return
for arranging and managing the loan, the lead arranger(s) receives a fee from the
borrowing ﬁrm. Through the relationship established with the borrowing ﬁrm, the
lead arranger is in a privileged position to investigate and monitor it. However, its

costly monitoring effort cannot be observed by the other members of the syndicate.

3

Therefore, to ascertain its commitment to monitor and induce other lenders to join the
syndication, the lead arranger must retain a large stake in the loan as only by putting
its own money at risk it will have an interest in the performance of the borrower
(Holmstrom and Tirole, 1997; Suﬁ, 2007). Thus, the structure of a syndicated loan
- meant as the loan share of the lead arranger(s) and the concentration of the loan -
offers rich information on creditors’ monitoring incentives.2

The empirical results reveal that, controlling for a variety of ﬁrm and contract char-
acteristics and for aggregate effects, the concentration of syndicated loans granted to
chaebol ﬁrms was lower than that of loans granted to non-chaebol ones and that after
the chaebol reform the difference in loan concentration narrowed. This is consistent
with our null hypothesis that the safety net generated by the government bailout pol-
icy and the chaebol cross-debt guarantees as well as the lack of accurate accounting
information diluted lenders’ incentive to monitor chaebol ﬁrms. When the reform
dismantled the safety net and fostered chaebol transparency lenders perceived the
need and the possibility to step up their monitoring effort and started to form more
concentrated syndicates to provide lead arrangers with monitoring incentives. These
results are conﬁrmed when we employ an alternative measure of monitoring intensity,
the participation of foreign lenders to the arrangement of syndicated deals. In emerg-
ing economies, foreign banks are allegedly tougher monitors than local banks because
they can rely on more accurate assessment techniques and are less subject to political
pressures (Kroszner, 1998; Giannetti and Ongena, 2008). We then delve deeper into
the data to disentangle the channels through which the institutional environment af-
fected creditors’ monitoring. First of all, the results are primarily driven by the top
ﬁve chaebols and only to a lesser degree by smaller, less inﬂuential chaebols. Since the
“big ﬁve” were the chaebols most protected by the bailout policy, this points to a role

of the government bailout policy in shaping creditors’ monitoring. Second, we ﬁnd

 

2Syndicated loans are a very important source of ﬁrm ﬁnancing in emerging and developed
economies. Later in the analysis, we shall provide evidence on this for the Asia-Paciﬁc region.

4

that the larger the equity stake of domestic banks in chaebol ﬁrms, the lower lenders’
monitoring and that this effect faded after the reform. This matches the idea that
state controlled domestic banks acted as a channel for bailout execution: until the
bailout policy was dismantled, their presence in the ranks of shareholders reassured
creditors that the government would intervene in the event of borrower distress.

In the ﬁrst part of the empirical analysis, we treat the agency problems that lenders’
monitoring can mitigate as a black box. In the second part, we take a preliminary
step towards investigating this issue and examine the interaction between creditors
and shareholders. We obtain some evidence that in ﬁrms with worse incentives for
controlling shareholders - as reﬂected in a larger gap between control and cash flow
rights - creditors’ monitoring was stronger. When the reform improved internal gov-
ernance mechanisms and the accountability of controlling shareholders, this effect
weakened, and especially so for chaebol ﬁrms.

The rest of the paper unfolds as follows. In Section 2, we relate the analysis to the
prior literature. Section 3 describes the institutional environment. Section 4 outlines
testable hypotheses on the impact of chaebol affiliation and reform on the structure
of syndicated loans. In Section 5, we provide details on the data. Section 6 presents
the core tests and results. In this section, we also argue that our results cannot be
rationalized by competing theories on the determinants of syndicate structure (such as
the diversiﬁcation motive theory). In Section 7, we subject the results to a variety of
robustness exercises. In this section, we also extend the empirical analysis to evaluate
the impact of chaebol institutional arrangements on lending standards, such as the

size, maturity and collateralization of syndicated deals. Section 8 concludes.

1 .2 Prior Literature

This study relates to two strands of empirical literature. The ﬁrst strand investigates
the determinants of ﬁnancial contracts. In this strand, the most closely related papers
examine lenders’ monitoring incentives in the context of syndicated loans. Employ-
ing respectively data on syndicated loans to US. and UK. ﬁrms, Suﬁ (2007) and
Bosch and Steffen (2008) ﬁnd that ﬁrms’ informational opaqueness shapes the struc-
ture of debt agreements. While insightful for examining such issues, these studies on
Anglo—Saxon economies are of limited help for discerning how institutions prevalent
in emerging economies, such as government bailout policies, affect monitoring and
discipline in the credit market (Vives, 2006; Schneider and Tornell, 2004).3 Another
set of studies in this literature explore the impact of laws and institutions on ﬁnan-
cial contracts. La Porta, Lopez-de—Silanes, Shleifer, and Vishny (1997) demonstrate
that countries with better investor protection and legal enforcement have larger and
more efficient capital markets. Focusing on equity holdings, La Porta, Lopez-de-
Silanes, Shleifer, and Vishny (1998) ﬁnd that stronger investor protection encourages
investors to hold smaller equity positions. Qian and Strahan (2007) document the
role of country-level legal factors, especially creditor protection, in shaping credit con-
tract terms, such as interest rates and loan maturity. Esty and Megginson (2003) use
data on project ﬁnance loans and investigate how cross-country differences in creditor
rights and legal enforcement affect the structure of syndicates. These studies focus
on a set of institutions different from ours. In our country study, in fact, we primarily
seek to understand whether, for a given court system and legal framework, the policies
of protection of business groups set up by the government (such as bailouts and fa-

vorable regulatory treatment) distort ﬁnancial contracting and creditors’ monitoring.

 

3Rajan and Zingales (2003) stress the importance of transparency for investors and creditors.
La Porta et al. (1998) argue that accounting standards play a critical role in corporate governance
by informing investors and rendering contracts more veriﬁable. Baek, Kang, and Park (2004) ﬁnd
that during the east asian crisis higher disclosure quality was associated with better stock price
performance. 6

Furthermore, a country analysis, in conjunction with the natural experiment offered
by the chaebol reform of the late nineties, can help mitigate the measurement and
endogeneity issues that affect cross-country empirical settings. For example, in such
settings it is hard to control for country speciﬁc laws that can affect credit market
decisions.

The second related line of research investigates the performance of business groups
in ﬁnancial markets. Bae, Kang, and Kim (2002) test whether chaebol ﬁrms beneﬁt
from acquisitions or whether such acquisitions just allow controlling shareholders to
increase their wealth by increasing the value of other group members (“tunneling”).
Hoshy, Kashyap and Scharfstein (1990) present evidence that the investment of ﬁrms
affiliated to Japanese business groups (keiretsu) is less sensitive to cash ﬂow than that
of stand alone ﬁrms. They interpret this as evidence that the monitoring performed by
main banks within keiretsu mitigates information problems. Business conglomerates
of developing countries are an interesting object of research because governments’
industrial policy guarantees them a special treatment in the event of distress. A
target of our paper is to understand the implications that this has for the monitoring
role of the credit market. Moreover, while until recently bailouts have been sporadic
in developed economies, during the 2008—2009 crisis the governments of the United
States and of other developed countries have rescued several business groups. Thus,
our study can also yield insights into the consequences that these bailouts will have

on the discipline role of the credit market of advanced economies.

1.3 Institutional Background

This section describes the institutional arrangements of chaebols that can shape cred-

itors’ monitoring incentives and the structure of credit contracts.

The Pre—reform Scenario

In South Korea, big business groups, chaebols, are owned and controlled by the
founder or his family successor, essentially operated as a single ﬁrm, and typically
aligned with the government (Choi and Cowing, 1999).4 More precisely, the South
Korean antitrust regulator, the Korean Fair Trade Commission (KFTC), annually
deﬁnes chaebol afﬁliates as those for which “either one person, his relatives, or a
company controlled by him own more than 30% of issued shares or substantially
affect the management (such as appointing its ofﬁcers) (Chang, 2003)”. A chaebol
comprises many subsidiary ﬁrms (typically around 30-50) which operate in different
industries under the same name (e.g., Samsung, LG Electronics, Hyundai Motor Co.).
Based on the deﬁnition of the Korean Fair Trade Commission, the literature agrees
in deﬁning “chaebols” the top 30 conglomerates (Youngmo, 1999; Jung, Kim and
Kim, 2008; Hwang, Park and Park, 2008). In addition, “chaebols” have frequently
been classiﬁed into two categories: the “big ﬁve” (Tierl henceforth), that is the ﬁve
chaebols with the largest book asset value, and the smaller chaebols (Tier2 henceforth)
(Youngmo, 1999). In fact, as we elaborate below, the government treatment of these
two classes of chaebols has been quite different, with the bailout policy primarily
protecting the top ﬁve chaebols (in line with the “too big to fail doctrine”).
Chaebols were created during the military dictatorship of the 19608. In the 19703,
they became the key partners of the South Korean government in promoting economic
development. The government, a major stockholder in national commercial banks,
channelled banks’ credit to chaebol ﬁrms to promote strategic sectors (e.g., heavy,
chemical, and export—driven industries). Throughout the 19808 and 19903, chaebols
further strengthened their role in the Korean economy and expanded their activity

into a broad range of industries, from commodities to high tech, from manufacturing

 

4In 1995, chaebols accounted for about 16% of South Korean GDP, 41% of manufacturing GDP,
5% of employment, 50% of exports, and 14% of total commercial bank loans. The top ﬁve chaebols
alone accounted for roughly 10% of GDP, over 40% of exports and a third of manufacturing GDP.

8

to the third sector.

Two institutional arrangements of chaebols are especially relevant for understand-
ing creditors’ monitoring incentives. The ﬁrst is the safety net that protected them.
To shield chaebols from the risk of failure, the Korean government supported them
with an implicit bailout policy (Lim, Haggard, and Kim, 2003). When a chaebol was
in distress, the government intervened by asking domestic, state-controlled banks to
write off bad loans.5 Large chaebol ﬁrms in distress received subsidized loans and
capital injections through domestic banks during the 1972 debt crisis (see the Au-
gust 1972 Emergency Decree), the 1979-1981 restructuring of heavy and chemical
industries triggered by the second oil shock, and the 1984-1988 wave of business in-
solvencies. Even in the absence of a bailout, chaebol ﬁrms could lobby the Bank
Supervisory Board (formerly the Financial Supervisory Commission) for a favorable
treatment by creditors (Chiu and Job, 2004). The government bailout policy was
compounded by the network of cross-debt payment guarantees among chaebol sub—
sidiaries, whereby members of a chaebol used their equity to secure the loans granted
to other members of the same chaebol. A second key characteristic of chaebols was
the lack of reliable accounting information on them. Besides having access to inter-
nal capital markets, thanks to the support of the government, chaebol ﬁrms could
easily obtain bank loans.6 Hence, they did not need to release accurate information
to outside investors to attract ﬁnancing. The interlocked share ownerships and re-
lated party transactions among chaebol ﬁrms further inhibited lenders’ acquisition of

information.

 

5111 1984, the government let a nonviable chaebol (Kukjae) go bankrupt but this was an isolated
episode.

6Shin and Park ( 1999) ﬁnd that the investment of chaebol ﬁrms is signiﬁcantly affected by the
cash ﬂow of other members of their chaebol.

The Post-reform Scenario

In the second part of 1997 and the ﬁrst part of 1998, a severe, albeit relatively short,
crisis hit South Korea. When the crisis broke out, the South Korean government asked
the IMF for an emergency bailout loan. The IMF provided assistance conditional on a
reform of the Korean economy and of the chaebols that restored market discipline. A
ﬁrst step of the reform consisted of dismantling the safety net that protected chaebols.
In 1997, the government let six chaebols (Hanbo, Sammi, J inro, KIA, Haitai, and New
Core) go bankrupt, a sort followed in 1999 by the second biggest chaebol, Dacwoo.
Moreover, the biggest chaebol, Hyundai, was split into Hyundai Motor Vehicle Co.
and Hyundai group while the sixth largest chaebol, Ssangyong, and several chaebol
affiliates engaged in bank-led workout programs. The decision of the government to
let some chaebols go bankrupt or undergo massive restructurings was compounded by
the introduction of tougher rules for corporate reorganizations. The amendments to
Korea’s Commercial Code (sang-bup) introduced in February 1998 established strict
time limits for in-court reorganizations and strengthened the power of creditors and
the reliance on experts in such reorganizations (OECD, 2000). In conjunction with
the lifting of the government bailout policy, the reform also marked the progressive
resolution of cross-debt guarantees among chaebol subsidiaries from 1998 to 2000.7
A second pillar of the reform of chaebols was the improvement in their accounting
transparency. To foster the ability of investors to monitor them, beginning in ﬁscal
year 1999 the Korean Financial Supervisory Service requested chaebols to report com-
bined ﬁnancial statements of all afﬁliated ﬁrms (kyulhapjaemujepyo system) rather
than consolidated ﬁnancial statements and to follow international accounting prin-

ciples, including quarterly reporting.8 Moreover, external auditors and accounting

 

7The amount of loans with guarantee was 26.9 trillion won in April 1998, 9.8 in April 1999, 4.3 in
December 1999, zero in March 2000 (Shin and Chang, 2003).

8A combined ﬁnancial statement brings together assets, liabilities, net worth, and operating
ﬁgures of two or more companies afﬁliated to a chaebol. Unlike consolidated ﬁnancial statements
(that include ﬁnancial information on subsidiaries only), combined financial statements cover all

10

ofﬁcers were henceforth to face harsher punishments if they misreported or falsiﬁed

ﬁnancial statements.9

1.4 Theoretical Hypotheses

In this section, we discuss the theoretical hypotheses. We start with the baseline
hypotheses on the impact of chaebol afﬁliation and chaebol reform on the structure

of syndicated loans.

1.4.1 Baseline Hypotheses

Scholars and policy-makers conjecture that the (implicit or explicit) safety net of-
fered by the government bailout policy and the chaebol cross-debt guarantees diluted
lenders’ incentive to monitor and discipline chaebol ﬁrms (see, e.g., Vives, 2006;
Schneider and Tornell, 2004; Lim, Haggard, and Kim, 2003; Nam, Kang, and Kim,
1999, and references therein). Creditors’ monitoring of chaebols would have been fur-
ther deterred by the opaqueness of the accounts and the lack of adequate disclosures
(Min, 2007). In turn, according to many analyses, the poor discipline imposed by
creditors was one of the causes of the low proﬁtability of chaebol ﬁrms during the
early 19903. After 1997, the scenario radically changed. The government repeat-
edly restrained from bailing out chaebol ﬁrms and removed cross-debt guarantees so
that henceforth creditors should have had stronger incentives to monitor. In addition,

creditors now had better tools to monitor chaebol ﬁrms, thanks to the introduction of

 

ﬁrms within a chaebol that are de facto under the control of the same shareholders (Chang and
Shin, 2007). ‘

9The chaebol reform carried out in concert with IMF/IBRD was known as the “Five plus Three
Principles” (Haggard, Lim, and Kim, 2003). Other objectives of the reform were (1) to reform
chaebol ﬁnancial structures, by reducing debt-equity ratios to 200% by the end of 1999; (2) to
establish core competencies of chaebol, for example swapping ﬁrms/ factories among them; (3) to
increase accountability, through the appointment of outside directors to the management boards of
the chaebol and through the strengthening of the rights of minority shareholders; and (4) to prohibit
interlocking share holdings among group members.

11

chaebol combined ﬁnancial statements and more demanding accounting requirements.

Building on the theoretical analysis in Holmstrom and Tirole (1997) and the em—
pirical analysis in Suﬁ (2007), we use the concentration of syndicated loans as a proxy
for creditors’ monitoring incentives. Holmstrom and Tirole (1997) show that a lender
has a stronger incentive to monitor a borrowing ﬁrm when it invests more of its own
money in the ﬁrm and, hence, has more to lose in the event of its default. Applying
this intuition, Suﬁ (2007) examines the impact of ﬁrms’ informational transparency
on creditors’ monitoring by looking at the structure of syndicated loans (e.g., loan
share held by lead arrangers or the concentration of the loan). Using loan concen-
tration as a proxy for lenders’ monitoring incentives, we can then state the following

hypothesis.

Hypothesis 1: The safety net that protected chaebols and the lack of accounting in-
formation on them diluted creditors' incentive to monitor chaebol firms. The loans to
chaebol affiliated firms were less concentrated than those to non-chaebol firms but after

the reform the difference in loan concentration narrowed.

As predicted by the “too big to fail doctrine”, the bailout policy of the South Korean
government especially protected the top ﬁve chaebols, which were viewed as being
of systemic importance for the Korean corporate sector (OECD, 2000). Therefore, if
the safety net generated by the bailout policy had a role in creditors’ monitoring, less
intense monitoring was needed for the “big ﬁve” chaebols but the need for monitoring

them rose disproportionately after the reform.

Hypothesis 2: If the safety net generated by the government bailout policy affected
creditors' monitoring, Tierl chaebol firms were granted less concentrated loans than other

firms. After the reform the difference in loan concentration narrowed.

Although our preferred approach to studying creditors’ monitoring incentives fo-

cuses on the syndicate structure, we can complement this approach with a second test

12

that looks at the composition - rather than the structure - of syndicates. Multinational
banks operating in emerging economies can allegedly count on better technologies and
personnel than local banks for processing “hard” information on borrowers, such as
accounting information (Detragiache, Tressel, and Gupta, 2008). Moreover, unlike
local banks, foreign ones are less connected with ﬁrms’ management and less subject
to political pressures, which can make them more independent monitors (Kroszner,
1998; Giannetti and Ongena, 2008).

Based on these arguments, we expect that the presence of foreign lead arrangers in a
syndicated loan signals more rigorous monitoring. We can thus revisit the predictions
on creditors’ monitoring using the presence of foreign lead arrangers rather than the

syndicate concentration as a proxy for its intensity.

Hypothesis 3: Foreign banks acted as lead arrangers less for chaebol firms than for
non-chaebol ones. After the reform their role in arranging loans for the two groups of

firms became more similar.

1.4.2 Other Hypotheses

We now turn to the channels through which we expect that chaebol institutional
arrangements impacted syndicated deals. Faccio, Masulis, and McConnell (2006)
argue that when state controlled institutions hold equity positions in distressed ﬁrms
the government has a stronger incentives to bail them out. Indeed, the domestic
banks most exposed to the default of the borrower could pressure the government
for a bailout. In addition, the presence of such institutions in the ranks of the ﬁrm’s
shareholders would ease the execution of a bailout. In fact, when a chaebol ﬁrm was
on the verge of bankruptcy, the government could bail it out through the domestic
banks that were stakeholders of the ﬁrm in question (see, e. g., the rescue package in
the 1972 Emergency Decree). We then expect that before the reform the presence

of domestic banks among chaebol ﬁrms’lghareholders reassured creditors that the

government would intervene in the event of borrower distress. Therefore, if the safety
net generated by the government bailout policy affected creditors’ monitoring, less
intense monitoring was needed for these ﬁrms, but the need for monitoring such ﬁrms

rose after the reform.

Hypothesis 4: If the safety net generated by the government bailout policy affected
creditors' monitoring, chaebol firms in which domestic banks had a larger equity stake
were granted less concentrated loans than other firms. After the reform, the gap in loan

concentration narrowed.

Together with the government bailout policy, the second pillar of the safety net that
protected chaebols was the system of cross-debt payment guarantees among chaebol

affiliates. This leads to the next hypothesis.

Hypothesis 5: If the safety net generated by chaebol cross-debt guarantees affected
creditors' monitoring, firms that could count on more support by their chaebols were
granted less concentrated loans. After the reform, loan concentration increased relatively

more for these firms.

We will discuss proxies for the ﬁnancial protection offered by chaebol cross-debt
guarantees later in the analysis. Finally, a last important point is that internal gover-
nance controls and creditors’ monitoring can act as substitutes in ameliorating agency
problems within ﬁrms. We thus expect that ﬁrms with weaker internal governance
were monitored less intensively by creditors and that the interplay between credi-
tors’ monitoring and internal governance controls changed after the reform. We shall

elaborate on these issues shortly.

1.5 Data and Measurement

This section provides details on the data sources, the measurement of the variables,

and the properties of the sample. Before 1(ielving into the data, however, it is useful

to form an idea about the Korean syndicated loan market.

1.5.1 Preliminary Observations

Figure I plots the total volume of syndicated loans granted in South Korea in each
year between 1992 to 2007 (the data are from various issues of the Securities Statistics
and Syndicated Loans of the Bank for International Settlements). In the left panel,
we compare it with the volume of syndicated loans in the whole Asian Paciﬁc region,
in the right panel with the volume of syndicated loans in the single countries of
the regibn. The Korean syndicated loan market was very active from 1994 to 1997
(averaging about eight billion US. dollars per year). When the crisis occurred in
1997, the volume of syndicated loans dropped dramatically. After the crisis, the
market began to expand again and in 2005 it reached the pre—crisis level. In 2007, the
volume of loans experienced a considerable rise (about 70%) relative to the previous
year level. As Figure I makes clear, the pattern of activity of the Korean syndicated
loan market is similar to that of the other countries of the region, though on a larger

scale.

1.5.2 Data Description

We draw the data for our empirical analysis from six sources: the Loan Pricing Corpo-
ration’s DealScan database compiled by Reuters; the KIS-Line data providing system
operated by Korea Investors Service Inc., a major Korean credit rating company; the
database on chaebol afﬁliation of the Korean Fair Trade Commission (KF TC); the
analysis of chaebols by Lee (2005); the DART system, an electronic disclosure system
run by the Korean Financial Supervision Service; and the Compustat Global Finan-

cial Services ﬁle. We obtain our sample of syndicated loans from DealScan,10which

 

lDealScan registers loans at the origination. League tables ranking the major providers of loans
are used for marketing purposes in the syndicated loans market. Therefore, as also noted by Ivashina
(2008), it is in the interest of lenders to report tligse data.

contains detailed information on terms, lead arrangers, and participant lenders of the
syndicated contracts. The original sample includes 4,692 syndicated loans granted
from July 1990 through September 2007 to ﬁrms located in Korea. For the 1990-1992
period there are only a small number of loans (16) in the dataset and for some of
these loans we lack information on their characteristics. Thus, we drop these observa—
tions. After excluding ﬁnancial ﬁrms (SIC codes 60-64) and government agencies (SIC
codes 91-99), there remain 2,686 loans. We also exclude deals that do not have lead
arrangers, bilateral loans, and loans granted to ﬁrms not identiﬁed in the KIS—Line
dataset. We then use the KIS-Line data bank to match loans with the characteristics
of borrowing ﬁrms (except for data on ﬁrms’ ownership structure which we obtain
separately from the staff of KIS—Line). KIS-Line is the largest and most comprehen-
sive data bank in Korea in terms of information on ﬁrms (including their corporate
ﬁnance and governance) and business groups. We hand-match the syndicated loans
in DealScan with the ﬁrm-level data in KIS-Line based on ﬁrm names and industry
classiﬁcation codes. A problem we face in this matching procedure is that in the
aftermath of the crisis a number of ﬁrms went through various kinds of restructur-
ing and, in the process, some changed name. DealScan sometimes refers to the old
names of' borrowing ﬁrms so that we have to trace the history of ﬁrms that changed
name. For this purpose, we checked history proﬁles on ﬁrms’ websites. Usually, the
websites report the date in which a ﬁrm changed name and information about when
it merged or was split (if it did). To double-check the identiﬁcation and obtain some
missing information, we integrate the websites with the KIS-Line’s ﬁrm proﬁle and
the DART system, which allows companies to submit disclosures online. The ﬁrms for
which neither the websites nor KIS-Line’s ﬁrm proﬁle or the DART system provide
enough information for identiﬁcation are dropped from the sample. After the whole
data cleanup, we are left with 1,626 loans granted to 242 non—ﬁnancial ﬁrms.

In order to identify chaebol afﬁliation, we use the KFTC database and the analysis
16

by Lee (2005). The KF TC database presents the ﬁrms affiliated to chaebols from
2001. For the years prior to 2001, we identify chaebol afﬁliated ﬁrms by referring
to Lee (2005). Furthermore, for the purposes of this paper, we do not classify the
privatized POSCO, KT&G and KT groups as chaebols because, although since 2001
they have been classiﬁed as such, their equity ownership is dispersed and no individual
person or family controls the conglomerate. As anticipated, in our analysis we classify
the top ﬁve chaebols for book value of assets as Tierll chaebols and the lower ranked
chaebols as Tier2 chaebols. The key reason for this distinction is that the government
bailout policy especially protected the “big ﬁve” (see, e.g., OECD, 2000). Thus, we
expect that the effect of the bailout policy on the structure of credit contracts was
especially strong for ﬁrms afﬁliated to these chaebols.

In the robustness analysis, we also employ lender speciﬁc variables. For the data on
domestic banks, we rely on KIS-Line (with the exception of the Korean Development
Bank for which we resort to Compustat ﬁles because KIS—Line does not provide
information for the early 19903). For the data on foreign banks, we instead resort to

the Compustat Global Financial Services ﬁle.

1 .5.3 Measurement

A critical step in the empirical analysis is the identiﬁcation of the lead arrangers of the
syndicated loans. We use two approaches. The ﬁrst is borrowed from Ivashina (2009)
who, following in turn the Standard & Poor’s deﬁnition, identiﬁes lenders using the
titles in a syndicated loan. The administrative agent is deﬁned as the lead arranger.
If the syndication does not have an administrative agent, the lender(s) acting as
book runner, lead arranger, lead bank, lead manager, agent, or arranger is (are)
deﬁned as the lead arranger(s). The second approach relies on Suﬁ (2007). DealScan’s
custom report feature lists two lender categories under the headings of “Lenders-Lead

Arranger” and “Lenders—All Lenders”. Following Suﬁ (2007), we thus treat lenders
17

as lead arrangers if they are classiﬁed under the heading of “Lenders-Lead Arranger”
and as participants if they are classiﬁed under the heading of “Lenders-All Lenders”
but not under the heading of “Lenders-Lead Arranger”. In this paper, we mainly rely
on the approach of Ivashina (2009) while we use the categorization of Suﬁ (2007) as a
robustness check. The results obtained with the two measures are virtually identical.

To measure the concentration of a loan, we construct two variables: the percentage
share of the loan held by the lead arranger(s) (hereafter, also lead share); and the
Herﬁndahl-Hirschmann index (herﬁndahl), calculated using the shares of the loan of
the syndicate members. The Herﬁndahl index is computed as the sum of the squared
individual shares of the loan and ranges from zero to 10,000, with 10,000 being its
value when a lender retains 100% of the loan. In our sample, the correlation between
the two measures of concentration is 0.9665. In some tests, we also experiment with
the number of lead arrangers and the number of participant lenders as further (less
precise) proxies for loan concentration. The dummy variables chaebol, Tier] chaebol
and Tier2 chaebol take on the value of one when a ﬁrm belongs to a chaebol, a Tierl
chaebol, and a Tier2 chaebol respectively, and zero otherwise. To capture the effect
of the chaebol reform, we construct a binary variable, reform, which takes on the
value of one if the loan was originated after 1997 and zero if it was originated prior to
1998. Scholars and policy-makers agree in identifying 1998 as the year in which the
government laid out and implemented the key steps of the reform (see, e.g., OECD,
2000, or Krueger and Y00, 2002). However, the reader could remain concerned that
some aspects of the reform were carried out in stages between 1998 and 2000 and
that the change in the institutional environment could have had lagged effects on
lenders’ behavior. For this reason, in the robustness section we also report results
obtained allowing for a transition between the pre-reform and the post-reform period.
The variable foreign arranger is a dummy taking on the value of one if at least one

of the lead arrangers of the loan is foreign, and zero otherwise. In constructing this

18

variable we have to account for the fact that some foreign banks (e. g., SC First Bank
and Korea Exchange Bank) had Operated as domestic banks before being acquired
by foreigners. We classify suCh banks as (domestic because their management - and,
allegedly, their practices - did not change signiﬁcantly after the transfer of ownership.
The variable domestic bank equity measures the percentage of equity held by domestic
banks in the ﬁrm. Chaebol leverage is instead the ratio of the total sum of liabilities
to the total sum of assets of a chaebol afﬁliates.

Turning to ﬁrm speciﬁc characteristics, the literature has identiﬁed three main
determinants of creditors’ monitoring. The ﬁrst is the amount of public information
available on the ﬁrm. The larger the amount of public information, the lower the
need for lenders to monitor the business. Following prior studies (e.g., Petersen and
Rajan, 1994), we use two proxies for the informational transparency of a ﬁrm, its age
and its size. Old ﬁrms have a better established track record so that more information
regarding them is publicly available; large ﬁrms receive more attention by analysts
and the ﬁnancial press. We compute the age of a ﬁrm as the number of days from
its inception to the loan origination date. We use various measures of ﬁrm size. We
experiment with the total assets and gross sales of the ﬁrm (both expressed in trillions
of South Korean won) and with the number of employees in the year in which the
loan starts. The results are uniform across the three measures and, henceforth, we
concentrate on total assets. The literature also underscores the importance of the
ex-ante quality and riskiness of a ﬁrm in determining lenders’ monitoring: the lower
the quality and the higher the riskiness, the more intense monitoring needs to be. We
measure quality with the proﬁtability of the ﬁrm, as given by the (net) income/total
assets ratio. We instead measure riskiness with the leverage ratio (total liabilities
to total assets): leverage is often employed to forecast the probability of default
of a ﬁrm. Finally, the literature emphasizes that the higher the value of the liquid,

pledgeable returns of a ﬁrm, the lower the need for creditors to monitor it (Holmstrom

19

and Tirole, 1997). We thus insert the ratio working capital/total assets to capture
liquidity, where working capital is computed as the difference between current assets
and current liabilities. Moreover, we enlist the ratio tangible assets/total assets to
control for asset tangibility. Another important aspect we are interested in capturing
in the empirical tests is the ownership structure of a ﬁrm. KIS-Line lists the names
of ﬁrm shareholders, their percentage equity shares, and their relationships to the
ﬁrm. Based on this information, we construct the two measures proposed by Job
(2003), block ownership concentration and control ownership gap. More details on
these measures will be provided later in the analysis. We also include one-digit
sector dummies to control for industry attributes and year dummies to control for
macroeconomic effects.

Finally, in the empirical analysis we employ a number of contract speciﬁc charac-
teristics. These include the size of the loan (expressed in millions of US. dollars), its
maturity (expressed in days), as well as whether the loan is secured or not. We also

insert dummies for the purpose of the loan.11

1.5.4 Sample Properties

Table 1.1 presents sample summary statistics broken down into loan, ﬁrm and chaebol
characteristics. Our dataset records more loans originated after the reform than before
it (1027 versus 591 if we treat 1998 as the ﬁrst year of the post-reform period). This
stems from the differences in the number of years and in the volume of activity of the
Korean syndicated loan market between the two periods (as Figure I makes evident).
Tierl chaebol ﬁrms are the most active borrowers (637 deals) followed by non-chaebol
ones (581 deals) and Tier2 chaebol ﬁrms (400 deals). These ﬁgures reflect the size of

the ﬁve largest chaebols. The average loan counts about 6.5 lenders, 3.4 lead arrangers

 

1Possible purposes of a loan are: general corporate purposes/working capital, takeovers and
acquisitions, leveraged buyout, debt repayment/ capital expenditure, other.

20

and 3.1 participant lenders. Compared for instance with the US. syndicated loans
examined by Suﬁ (2007) for the 1992-2003 period, the number of lead arrangers is
twice as large whereas the number of participant lenders is almost half. For 1,023
loans, we know the shares held by the members of the syndicate. The average share
held by the lead arranger(s) is 25.8% and the average value of the Herﬁndahl index
for the loan shares is 3233. The statistics for a ﬁrm are computed as the average
across the loans granted to the ﬁrm. The mean age of a ﬁrm is about 22 years and its
average number of employees is 3,645. The average total assets amount to 2.5 trillion
South Korean won (1.98 billion US. dollars at the end-of-2000 exchange rate). The
mean leverage ratio is 0.63, a high ﬁgure compared with the mean leverage ratio of
0.34 of the US. ﬁrms examined by Suﬁ (2007). This reﬂects the large amount of
debt accumulated by Korean ﬁrms before the crisis (OECD, 2000). Turning to loan
characteristics, the most common loan purpose is general corporate purposes / working
capital followed by debt repayment / capital expenditure. The average size of a loan is
302 million dollars and the average maturity is about six months (the corresponding
ﬁgures for the US. loans investigated by Suﬁ, 2007, are 364 million dollars and about
three months). Roughly 22% of the loans are secured.

Table 1.2 also investigates differences between chaebol and non-chaebol afﬁliated
ﬁrms. Let us ﬁrst consider the structure of syndicated loans. Tierl chaebol ﬁrms
have the lowest loan concentration, whether this is measured by the total share kept
by the lead arranger(s) or the Herﬁndahl index. For Tierl and Tier2 chaebol ﬁrms
loan concentration signiﬁcantly rises from the pre- to the post-reform period while
for non-chaebol ﬁrms loan concentration is roughly the same in the two periods (the
Wilcoxon test is not signiﬁcant at meaningful levels). The average number of lenders
in syndicated loans to Tierl chaebol ﬁrms is 7.6 which is larger than that in loans
to Tier2 chaebol or non-chaebol ﬁrms (5.6 and 6 respectively). Such a difference

is more pronounced before the reform. The number of lead arrangers is the largest

21

for Tierl chaebol ﬁrms both on average and in each of the two sub-periods. This is
also true for the number of participant lenders, although for Tierl chaebol ﬁrms the
number of participants exhibits a decline after the reform. All in all, inspection of the
unconditional means suggests that loan concentration tends to be lower for chaebol
(especially Tierl chaebol) ﬁrms than for non-chaebol ones but that the difference
narrows after the chaebol reform. Turning to ﬁrm characteristics, Tier2 chaebol ﬁrms
are the oldest and non-chaebol ﬁrms are the youngest. Looking at employees and total
assets, Tierl chaebol ﬁrms are the largest while Tier2 chaebol and non-chaebol ﬁrms
have similar size. The leverage ratio does not vary much across the three groups but
it signiﬁcantly declines after the reform, especially for Tierl and Tier2 chaebol ﬁrms.12
Proﬁtability, proxied by the income to total assets ratio, does not differ between Tierl
and Tier2 chaebol ﬁrms while on average non-chaebol ﬁrms have a negative income
after the reform. The ratio of tangible assets to total assets is roughly 0.4 for all

groups and remains stable over the sample period.13

1.6 Empirical Tests and Results

In this section, we discuss the empirical models and our core ﬁndings. We start with
the baseline results and then try to disentangle the channels through which chaebols

impact the structure of syndicates.

1.6.1 Chaebols and Syndicate Structure

The baseline model we estimate is

Syndijt = a + 321+ lehaebolj + nghaebolj - Reform + de + Xzfé + Eijt' (1.1)

 

1After the reform, leverage converged to about 0.60 for all the three groups.
1Looking at loan characterisics, on average the loans to non-chaebol and Tier2 chaebol ﬁrms were
larger, of longer maturity, and more frequently secured than those to Tierl chaebol ﬁrms.

22

The dependent variable is a measure of the structure of syndicated loan 2' granted
to ﬁrm 3' in year t, such as the total share retained by the lead arranger(s) and
the Herﬁndahl—Hirschmann index for the loan shares. The key right-hand-side vari-
ables of interest are chaebol and its interaction term with reform: the coefﬁcients
7] and 72 measure how “chaebol afﬁliation” and “affiliation after the reform” affect
the structure of the loan. In the analysis, we also differentiate between Tierl and
Tier2 chaebol ﬁrms (replacing chaebol with Tier] chaebol and Tier2 chaebol). In the
richest speciﬁcation, the vector X j of ﬁrm-level control variables include (the natural
logarithm of) ﬁrm total assets, the leverage ratio, the ratio of income to total assets,
(the natural log of) ﬁrm age, the ratios of tangible assets to total assets and working
capital to total assets, and industry dummies. The vector X 2’ of contract-level control
variables comprise (the natural logs of ) loan maturity and loan amount, and dummies
for the loan purpose and for whether the loan is secured or not. We also insert the
time (year) dummies [it to control for macroeconomic conditions that affect all ﬁrms.
Finally, throughout the analysis, we use robust standard errors clustered at the ﬁrm
level.

As discussed earlier, our main hypothesis is that the safety net generated by the
government bailout policy and the chaebol cross-debt guarantees, together with the
poor quality information available on chaebols, diluted lenders’ incentives to moni-
tor chaebol ﬁrms. Under this hypothesis, the impact of chaebol afﬁliation on loan
concentration - as captured by the coefﬁcient 71 - should be negative. As for the
coefﬁcient 72 on the interaction term, this reﬂects the impact that the chaebol reform
had on the structure of syndicated loans granted to chaebol ﬁrms and, hence, on lead
arrangers’ incentive to monitor such ﬁrms. We expect 72 to be positive, reﬂecting
a step-up in creditors’ monitoring effort after the safety net was dismantled and the
quality of accounting information on chaebols improved.

Table 1.3 displays the results of the baseline model. In Panel A regression (1),
23

we use the Herﬁndahl index as the dependent variable. The estimated coefﬁcient on
chaebol is negative while the coefﬁcient on the interaction term is positive, with the
coefﬁcients being statistically signiﬁcant at the 10% and 5% level, respectively. The
magnitude of the impact appears to be sizeable: the afﬁliation to a chaebol reduces
the Herﬁndahl index by 516.9, that is approximately 16% at the sample mean. When
we add a battery of ﬁrms’ characteristics, as in regressions (2)-(4), and loan attributes
(regression (4)) the signs and signiﬁcance of the coefﬁcients do not change (except in
regression (3), in which the coefﬁcient of chaebol becomes marginally insigniﬁcant).14
In Panel B, we use the loan share of the lead arrangers as a measure of loan concentra-
tion. The results are conﬁrmed: for example, in the most parsimonious speciﬁcation
of regression (5) chaebol afﬁliation results in a lower loan share of the lead arrangers
by approximately 22% at the sample mean. Overall, the estimates support our main
hypothesis that the safety net that protected chaebols and the lack of quality in-
formation deterred creditors’ monitoring of chaebol ﬁrms and that after the reform
creditors set up stronger monitoring incentives. The results for ﬁrm-level controls
are also in line with expectations. In regressions (2) and (3), the negative coefﬁcient
on (log)assets indicates that as ﬁrm size expands loan concentration declines. This
suggests that higher ﬁrm transparency, as reﬂected in a larger business size, requires
less intense monitoring and, hence, lower borrowing concentration. The coefﬁcient
on leverage is positive and statistically different from zero. As for the coefﬁcient on
income to total assets, this is negative and also signiﬁcant (at the 1% level). These
ﬁndings are consistent with the conjecture that worse ﬁrm ﬁnancial conditions, as re-
ﬂected in a higher leverage and a lower proﬁtability, call for more intense monitoring
and higher borrowing concentration. In Panels C and D of Table 1.3, we reestimate
the regressions differentiating between Tierl and Tier2 chaebol ﬁrms. The estimates

reveal that the effect of chaebol afﬁliation and of the policy reform on loan concen-

 

1Clearly, in regression (4) including loan characteristics may be the source of an endogeneity
problem.

24

tration are pronounced for Tierl ‘chaebol ﬁrms. For Tier2 chaebol ﬁrms, the impact
of chaebol afﬁliation is instead considerably weaker and seldom signiﬁcant. However,
after the reform the afﬁliation to a Tier2 chaebol appears to have a positive and
statistically signiﬁcant effect on loan concentration.

Panel A in Table 1.4 tabulates the point estimates obtained focusing on the number
of lead arrangers. The coefﬁcient 71 in regression (4) turns out positive and signif-
icantly distinct from zero at the 5% level of conﬁdence. The sign of 71 reads the
same way as in the regressions in Table 1.3. The regressions in Panel C offer some
evidence that Tierl chaebol afﬁliation positively affects the number of lead arrangers.
(while Tier2 chaebol afﬁliation appears to have no impact).15As columns (2)-(3) illus-
trate, the coefﬁcient on ﬁrm assets is statistically different from zero at the 5% level
and suggests that as ﬁrm size increases more lead arrangers join the deal. Size also
appears to explain most of the variation in the number of participant lenders.

The stronger economic and statistical signiﬁcance of Tierl chaebol afﬁliation and
reform points to a role of the government bailout policy in shaping creditors’ moni-
toring incentives. In fact, the bailout policy targeted the “big ﬁve” chaebols because
of their systemic importance in production and employment (OECD, 2000; Nahm,
2003; Jang, 2001). The ﬁnding that the effect for Tier2 chaebol ﬁrms is weaker can
signal that for smaller chaebols there were lower expectations of bailouts and, hence,

lenders’ monitoring was somewhat stronger (and, in turn, the effect of the reform

weaker).

1.6.2 Chaebols and Foreign Banks

It is often argued that in emerging economies foreign banks are less subject to po-

litical pressures and, hence, are tougher monitors than local banks (Kroszner, 1998;

 

1In an unreported table, we also reestimate the regressions by constraining the coefficients on
Tier2 chaebol and non-chaebol ﬁrms to be the same (thus, in these estimations Tier2 chaebol and
non-chaebol ﬁrms are the omitted group). The results carry through.

25

Giannetti and Ongena, 2008). Moreover, internationally active ﬁnancial institutions
(e. g., multinational banks) employ better trained loan officers, a more advanced infor-
mation technology, and more accurate risk assessment techniques. Thus, if our main
hypothesis is valid we would expect that the role of foreign banks as lead arrangers
was weaker for chaebol than for non-chaebol ﬁrms and that it strengthened after the

reform. To test these predictions, we estimate the following regression

Pijt(f0’r€ign arrangerz' : 1) : <I)(a + 31: + dlChaebolj+ ( )
1.2

(lgChaebolj . Reform + X,-6 + ng + am),

where foreign arranger is a binary variable that takes on the value of one if the
lead arrangers of loan i are foreign banks, and zero otherwise,16while X j and X;
are the vectors of ﬁrm- and contract-level control variables already incorporated in
equation (1.1).

In Table 1.1, we report summary statistics for the involvement of foreign banks in
syndicate formation. The percentage of the loans that have foreign lead arrangers is

20%, with this ﬁgure being lower(15%) for Tierl chaebol ﬁrms before the reform.

In Table 1.5, we present the results obtained using a probit model (the results of
a linear probability model are essentially the same). In line with our hypothesis,
the regressions provide some evidence that chaebol affiliation reduces the probability
that foreign lenders act as lead arrangers. However after the reform, for Tierl chaebol

ﬁrms, the probability increased.

 

lgn individual foreign bank can play a role of the lead arrangers in the three different settings;
as a single lead arranger, or as one of them for the loans whose lead arrangers are all foreign banks,
or as one of them for the loans whose lead arrangers are a mixture of foreign and domestic banks.
The participation decision of foreign banks as a lead arranger in the third setting may be affected
by that which domestic banks become their counterparts. About 40% of the loans belonging to the
third case have Korea Development Bank, which is owned by the government, as one of the co—lead
arrangers. Apart from the asymmetric information problem, a bailout expectation may have affected
the foreign banks’ participation decision when one of the co—lead arrangers is a domestic bank which
is heavily controlled by the government. To be true to the theory, the loans whose lead arrangers
are all foreign banks are considered as the treatment group in our analysis.

26

As for the control variables, ﬁrm size and (to a lower extent) proﬁtability positively
affect the probability that a foreign bank is a lead arranger. The result for size perhaps
reﬂects the fact that foreign lenders are more skilled at monitoring large businesses.
For example, Detragiache, Tressel, and Gupta (2008) show that foreign banks are
relatively better at interpreting “hard” information than “soft” information, and big
ﬁrms typically have richer hard information (e.g., credit reports and ratings) than
small ones.

As a robustness check, we redeﬁne the binary variable foreign arranger with the
one which takes the value of one if the lead arranger is a single foreign bank, zero
otherwise.17The estimates - reported in Panel B of Table 1.5 - conﬁrm the previous-

ﬁndings.

1.6.3 The Role of the Safety Net

The tests that follow aim at further disentangling the role of the chaebol safety net

in inﬂuencing creditors’ monitoring.

Domestic Banks’ Equity and Political Connections

While Korean banks were privatized in 1981-1983, the Government inﬂuenced credit
allocation through the appointment of bank CEOs until 1993 and its approval of
bank CEO appointments until political democratization began in 1997 (Yoo, 1999;
Joh, 2001). Top bank managers often had, short-term appointments (around three
years), which meant that they only needed to prop up failing ﬁrms for a few years
to avoid loan defaults during their tenure (Chiu and Joh, 2003). The pressure on

the government to rescue troubled chaebol ﬁrms is likely to have been particularly

 

1A foreign bank will be more willing to arrange the loans with other foreign banks when its own
information on the ﬁrms is limited, and other banks’ information is complementary. If this is the
case, then the probability of the loans being arranged by a single foreign bank will be higher for the
ﬁrms on which information is more available.

27

strong when domestic banks controlled by the government held an equity stake in the
ﬁrms (Faccio, Masulis, and McConnell, 2006). Additionally, the main channel through
which the government could carry out bailouts was the intervention of domestic banks.
These arguments suggest that creditors’ expectation of a bailout should have been
higher - and, correspondingly, their incentive to monitor chaebol ﬁrms lower - when
domestic banks held a larger equity stake in the ﬁrms. To verify this hypothesis, in
Table 1.6 we reestimate the regressions by inserting the percentage equity share of
domestic banks, its interactions with the Tierl and Tier2 chaebol dummies, as well
as the triple interactions with the reform dummy.

The results shown in columns (1)-(3) strongly suggest that the equity stake of
domestic banks negatively impacts the concentration of syndicated loans granted
to chaebol ﬁrms before the reform. Especially for Tier2 chaebol, such impact was
pronounced as shown in columns (7)—(12). In the triple interaction for Tierl chaebol
in columns (7)-(12), the loan concentration decreased after the reform.

This is in line with our hypothesis. A stronger presence of domestic banks in ﬁrm
ownership would have raised the chances of a government bailout, reducing the need
for monitoring.

However, after the reform the probability of a government bailout declined, shutting
down the channel through which domestic banks could support distressed ﬁrms. To
double check the presence of an “equity participation” channel, we consider other
entities whose equity participation may affect lenders’ monitoring incentives, such
as the central and provincial governments (while excluding domestic banks’ equity
shares). The results (non-tabulated to conserve space) are similar to those obtained
in the regressions where only domestic banks’ equity is considered: especially for Tier
2 chaebols the effect is pronounced in the pre—reform period but it vanishes after the
reform.

Following Faccio, Masulis and McConnell (2006), in additional tests in Table 1.11
28

we also consider the effect of chaebol informal political connections as a channel
alternative to domestic banks’ equity participation in offering protections to chaebol
ﬁrms. We construct the measure of political connection as a binary variable that takes
on the value of one if at least one of the equity holders of the chaebol (equity share
2 1%) was a member of the National Assembly of Korea in the year of origination of

the loan. However, the estimated coeﬂicient on this indicator is not signiﬁcant.

Intra—Chaebol Effects

The previous tests suggest that the government bailout policy shaped creditors’ mon-
itoring. We now attempt to isolate the role (if any) of the chaebol cross-debt guar-
antees, the second pillar of chaebol safety net. With this objective, we allow the
coefﬁcients on chaebol and the interaction term with the reform dummy to vary
according to key characteristics of the chaebol a ﬁrm belongs to (if any). Building on
Shleifer and Vishny (1992) and the related literature, our preferred way to capture
the role of cross-debt guarantees among chaebol afﬁliates is looking at the correlation
between the ﬁnancial conditions of the ﬁrm and those of the other chaebol members.
When this correlation is high the ﬁrm will tend to encounter distress when other ﬁrms
of the chaebol are also in trouble (i.e., the chaebol will have less diversiﬁed activities).
This will erode the protection offered by the chaebol cross-debt guarantees, requiring
more intense monitoring on the part of creditors. We compute the correlation be-
tween the net income to total assets ratio of the ﬁrm and that of the other chaebol
afﬁliates. The results (reported in Table 1.7, Panel C) indeed reveal that ﬁrms whose
income comoves more with that of their chaebol receive more concentrated loans but
that this effect weakened after the reform. A second, more simplistic way to capture
intra-chaebol effects is to look the impact of chaebol leverage. In Table 1.7, Panel
Band we ﬁnd a signiﬁcantly negative effect of chaebol leverage on loan concentra-

tion and the coefﬁcient on the interaction with the reform dummy indicates that

29

this effect weakened after the reform. One interpretation for the impact of chaebol
leverage is that this is picking up the positive impact that the debt capacity of the
chaebol has on the safety of the borrowing ﬁrm (thus, requiring lower monitoring on
the part of creditors). In turn, the coefﬁcient on the interaction term suggests that,
by removing the system of cross-debt guarantees, the chaebol reform may have led
lenders to pay less attention to the ﬁnancial conditions of the chaebols and more to
those of the individual borrowers. All in all, though non-conclusive, the ﬁndings thus
point to a role of chaebol cross-guarantees in impacting creditors’ monitoring: in the
pre-reform period, the ﬁnancial conditions of the chaebol would have been factored in
determining lenders’ monitoring effort; after the reform banned cross-debt payment

practices, this effect would have disappeared.

1.6.4 Internal and External Governance

Thus far, we have left the interaction between creditors and ﬁrm shareholders in the
background. We now turn to examine this interaction, possibly shedding light on
the corporate governance issues that creditors’ monitoring can ameliorate. In South
Korea, large shareholders (typically the ﬁrm founders or they families) control ﬁrms
even if they own little equity in them (Joh, 2003). This occurs because ownership is
dispersed and institutional investors do not scrutinize ﬁrms effectively. In the case
of chaebols, the divergence between control and ownership rights is exacerbated by
pyramid structures and cross-share holdings among ﬁrms. The corporate governance
literature stresses that when control rights are larger than ownership rights ineffi-
ciencies can arise because controlling shareholders beneﬁt from expropriating ﬁrm
resources (such as pursuing private beneﬁts at the expense of other stakeholders)

but suffer only in part from the costs of expropriation (Shleifer and Vishny, 1997).18

 

1The abuse of controlling shareholders may take various forms, such as cash ﬂow appropriation
or asset stripping (La Porta et al., 1998). Bae, Kang and Kim (2002) ﬁnd that while minority share-
holders of a chaebol afﬁliated ﬁrm making an acquisition lose, on average the controlling shareholder

30

By reducing the divergence between control and ownership rights, a concentrated
ownership should help align the incentives of large shareholders with those of other
stakeholders (Claessens, Djankov, Fan, and Lang, 2002).

On the ground of these arguments, one can conjecture that creditors’ monitoring
and the internal mechanisms of control of large shareholders, as proxied by an appro—
priate measure of ownership concentration, act as substitutes in disciplining ﬁrms. To
test this hypothesis, we use the two measures of internal governance proposed by Joh
(2003). The ﬁrst is the bloc ownership concentration, deﬁned as the sum of the largest
equity shares of the ﬁrm, excluding institutional shareholders (ﬁnancial institutions
and non-ﬁnancial corporations), foreign owners, government, and employment stock
ownership stakes. The reason we only include large personal holdings is that before
the 1999 revision of the South Korean Commercial Law institutional investors could
not exercise full voting rights (for instance, the “shadow voting” rule forbade ﬁnancial
institutions from voting on ﬁrm decisions).19As a second (inverse) metric of internal
governance controls we use the control ownership gap, deﬁned as the difference be-
tween the total equity holdings of the top shareholders and ownership concentration.
When constructing ownership concentration and control ownership gap, we exper-
iment both using the top ten and the top eight shareholders. The results in Ta-
ble 1.8 suggest that ownership concentration negatively affected loan concentration
(lead share or her findhal), with the coefﬁcients signiﬁcant at the 10% or 1% level
(see Panel A). Moreover, these effects weakened after the chaebol reform. The evi-
dence is however less compelling when we insert control ownership gap rather than
ownership concentration in the regressions: the coefficients are positive and statis-

tically signiﬁcant only when we consider the top eight shareholders (see Panel B).

 

beneﬁts because the acquisition enhances the value of other ﬁrms in the chaebol.

1Using a large sample of South Korean ﬁrms observed over the 1993-1997 period, Joh (2003)
ﬁnds that the largest shareholder of a publicly traded ﬁrm (and her family) owned about 31.7% of
shares on average while controlling shareholders of privately held ﬁrms owned 49.5% of shares on
average. Claessens et al. (2000) report that frequently the largest and controlling shareholder or
family members are among the top executives. 31

Turning to the role of chaebol afﬁliation before and after the reform, we obtain
that, after controlling for ownership concentration or control ownership gap, the
results for chaebol and its interaction with the reform dummy carry through (see
Table 1.9). Whether the stronger misalignment between control and ownership rights
inside chaebols had repercussions on creditors’ monitoring decisions is then tested by
interacting chaebol with the two governance indices.20The interactions between the
chaebol dummy and ownership concentration or control ownership gap are signiﬁ-
cantly negative (Table 1.10). This could suggest that the interplay between internal
and external governance controls is different for chaebol ﬁrms. Overall, though not
conclusive, the results suggest that creditors could need to monitor less intensely
ﬁrms in which ownership rights are better aligned with control rights. The ﬁnding
that after the reform the impact of ownership concentration on the structure of credit
contracts weakens could reﬂect the fact that the reform strengthened creditors’ rights
and mitigated corporate governance problems (Shin and Chang, 2003). In fact, the
South Korean government implemented several steps to curb the power of control-
ling shareholders, such as lifting the limitation on the voting rights of institutional
investors, introducing a system of class action, and improving the accountability of
directors (e.g., through the appointment of individuals from outside the companies
to the boards of directors). The reform thus rendered the rights of all stakeholders,

including creditors, better secured.

1.6.5 Monitoring or Diversiﬁcation?

We have interpreted the results in light of theories on creditors’ monitoring, such as
Holmstrom and Tirole (1997) and Diamond (1984). The main competing explanation

for the structure of syndicates hinges on the diversiﬁcation motive: lead arrangers

 

2The summary statistics in Table 1.2 reveal that on average Tierl chaebol ﬁrms have the lowest
ownership concentration and the largest control ownership gap.

32

could choose to syndicate out a loan to a larger number of participants to spread the
risk of default. For example, Qian and Strahan (2007) ﬁnd evidence of such a behavior
in countries where creditors’ protection is worse. We contend that our results cannot
be rationalized by the diversiﬁcation motive. In fact, the safety net generated by the
government bailout policy and the cross-debt guarantees mitigated the risk suffered
by lenders - indeed, we believe this is why creditors’ monitoring incentives should
have been diluted. If the diversiﬁcation story held, we would then expect that before
the reform lenders diversiﬁed loans less, and that the need for diversiﬁcation rose

after the safety net was dismantled. This is the opposite of what we ﬁnd in the data.

1 .7 Further Tests

We check the robustness of our results with various tests. We then extend the anal-
ysis to study the impact of chaebol institutional arrangements on lending standards,

namely loan size, maturity and collateral requirements.

1.7 .1 Robustness

We perform a variety of robustness tests (selected results are displayed in Table 1.11
while others are not tabulated to conserve space). First, we reestimate all the regres—
sions by partitioning the sample between the pre— and the post-reform period rather
than interacting the chaebol indicator variable with the reform dummy. The results
carry through (details available upon request). A second robustness check consists
of allowing for a transition phase between the pre- and the post-reform period to
account for the progressive implementation of some aspects of the chaebol reform.
The estimates obtained treating 1998 and 1999 as a transition period read the same
way as those discussed so far: the effect of the reform appears concentrated in the

post-transition period (see Table 1.11, Panel A). A third robustness exercise con-

33

sists of controlling for lenders’ characteristics. In fact, a growing body of theoretical
literature (e.g., Carletti, Cerasi and Daltung, 2007) predicts that lenders’ ﬁnancial
conditions could drive their monitoring decisions. We construct two variables, the
ratios cash/total assets and capital/total assets, to gauge the lead arrangers’ liq-
uidity and capitalization (if there is more than one lead arranger, here we treat the
one with the largest portion of the loan as the lead arranger). Because information
on these items is only available for a subset of our sample, in this speciﬁcation we
suffer a substantial loss of observations. The (untabulated) results are robust to the
inclusion of these variables; moreover, lenders’ liquidity and capitalization have no
signiﬁcant impact on loan concentration. Next, we reestimate all the regressions using
the deﬁnition of lead arranger(s) of Suﬁ (2007), obtaining virtually identical results.

A further robustness check consists of controlling for geographical location. In our
setting, this does not constitute a major issue as most economic activity in South
Korea is concentrated in the area of Seoul and economic conditions are not too dis-
similar across regions. Nevertheless, to assuage possible concerns, we experiment by
inserting a dummy that takes on the value of one if the ﬁrm is located in the capital
city area, and zero otherwise. The (unreported) results remain largely unaltered and
the geographical dummy turns out to be statistically insigniﬁcant. In a supplemen-
tary test, we experiment with an alternative measure of concentration proposed by
Harman (1997) which is computed by subtracting the inverse of the number of lenders
from the Herﬁndahl-Hirschmann index. This measure partly mitigates the impact of
the number of lenders on the Herﬁndahl index and, hence, better reﬂects loan share
asymmetries for a given number of lenders. The results carry through when we re-
place the Herﬁndahl index with this dependent variable (see Panel B of Table 1.11).
In an additional test (reported in Panel C of the table), we reestimate the baseline re-
gressions using a fractional response model. This is considered an alternative to OLS

when, as in our case, the dependent variable is a fractional response, that is, it falls

34

between zero and one. The results remain essentially unchanged with this approach.
In a last robustness check, we control for the reputation of the lead arrangers of the
loan. In fact, the incentive of a lead arranged to Shirk on her monitoring activity
could be stronger for banks that are less reputable. Following Suﬁ (2007), we proxy
a bank’s reputation with a dummy that takes on the value of one if the bank is one
of the top ten lead arrangers according to the league table for Korea compiled by
Dealogic and Thomson Financial. The inclusion of this variable does not alter the

results.

1.7 .2 Other Loan Characteristics

In this paper, we have primarily focused on the structure of syndicated loans. A
related research question that can enrich our understanding of lenders’ behavior is
whether the institutional environment affects other loan features. The DealScan
database provides information on three further attributes of syndicated loans: ma-
turity, size, and whether the loan is secured or not. The collateralization of loans is
a means alternative to monitoring by which creditors protect themselves from ﬁrms’
misbehavior (Holmstrom and Tirole, 1997). Diamond (2004) also contends that short-
term debt can usefully complement collateral and monitoring in ameliorating agency
problems. In fact, by generating a risk of “ﬁrm runs”, it can commit multiple lenders
to enforce claims, thus providing proper incentives to borrowers.

Table 1.12 reports the results obtained using maturity, size, and an indicator for
whether the loan is secured or not as dependent variables. The aﬂiliation of ﬁrms to
Tierl and Tier2 chaebols tends to shorten loans but the reform weakens this effect
for Tier2 chaebols (see columns 7 and 8). Moreover, leverage and the ratio of income
to total assets shorten the maturity of loans, while the tangible to total assets and
working capital to total assets ratios lengthen it. The latter two ﬁndings may signal

that ﬁrms with more collateralizable assets and more liquid returns can afford long-

35

term credit. Turning to loan size, the point estimates imply no significant differences
between chaebol ﬁrms and non-chaebol ones, and loan maturity for chaebols has not
changed after the reform either. Similarly, the extent to which loans are secured
is not affected by chaebol affiliation while its main determinants appear to be ﬁrm
size, leverage and asset tangibility. In particular, ﬁrms with larger assets seemingly
obtain loans with lower collateral requirements, possibly because of their higher in-
formational transparency. This is also the case for less levered businesses, probably
reﬂecting of the fact that ﬁrms less exposed to default risk need to post less collateral.

Overall, a possible interpretation of these additional tests is that the removal of
the bailout policy and of the chaebol cross-debt guarantees not only fostered lenders’
monitoring but also induced lenders to tighten lending standards, granting shorter-
term loans. Although we cannot test this conjecture, to the extent that before the
reform chaebol ﬁrms engaged in over-borrowing (see, e.g., Krueger, and Y00, 2002),
this interpretation could also point to higher selectivity on the part of creditors and

more efﬁcient investment policies after the reform.

1.8 Conclusion

This paper documents the impact of the institutional environment on the structure
of credit contracts using the South Korean syndicated loan market as an empirical
laboratory. As demonstrated by prior literature, the structure of a loan can yield
important insights into lenders’ effort in monitoring and disciplining borrowers. A
central result of the analysis is that, controlling for ﬁrm and contract speciﬁc charac-
teristics and for macroeconomic effects, chaebol ﬁrms tend to be granted less concen-
trated loans than non-chaebol ones and that this difference narrowed after the chaebol
reform enacted in the late nineties. These results suggest that prior to the reform

the safety net generated by the government bailout policy and the chaebol cross-debt

36

guarantees diluted creditors’ incentives to monitor chaebol ﬁrms. When the chaebol
reform removed bailout and cross-debt guarantees, creditors stepped up their mon-
itoring effort and toughened lending standards, granting smaller and shorter loans.
In further tests, we have explored the governance problem that creditors’ monitoring
is supposed to mitigate. We have obtained some evidence that lenders monitored
ﬁrms with better internal governance less intensively. Moreover, the reform of cor-
porate governance - and the consequent improvement in the internal discipline of
controlling shareholders - complemented the increase in creditors’ monitoring. We
have also found preliminary evidence that internal governance mechanisms interacted
with creditors’ monitoring differently between chaebol and non-chaebol ﬁrms.

The analysis leaves open interesting questions. In emerging economies, it is crucial
to understand how the reforms of different aspects of corporate governance interact
one with another. In this paper, we have made a ﬁrst step in this direction. How-
ever, more work is needed to understand the interplay among creditors, institutional
investors and minority shareholders in providing proper incentives for an efﬁcient
ﬁrm management. An investigation of this interplay between internal and external

governance mechanisms is the next objective of our research.

37

 

Figure 1 Volume of Syndicated Loans in Korea and Asia-Pacific Countries

 

140

120

100

 

—-Asia & Pacific

........... South Korea

92939495969798990001020304050607
Source : BIS, In billions of US dollars

 

 

35

30

25

20

 

 

 

--~".\

 

 

 

 

 

 

 

 

92 93 94 95 96 97 98

 

99 00 01 02 03 04 05 06 07
Source : BIS, In billions of US dollars

— Singapore
_ China

-— India

 

Indonesia

 

Malaysia

- '- - South Korea
-------- Taiwan, China
Thailand

 

 

38

 

 

.350 ~.— 9.35:3. M8338 8.. Mus—$8.2. rem: can—m

am 820 588:6 3:85 3538 8.. Ba 3568 on. {No @8323 8:: gnu—m 898335 NAN 333.. 3.0:: :5. $3 8
$6.382 Nooq. 9:338 won ma: 8a gnocc— 07888138 :8 cane—Boa 3 Ba an: as: :8 nuance. :22. 888385.. mszmaom
8a Ana 3888138 :8 cans—mag a 5» an:— :22.

 

 

2:8. 0cm. Zoe: mb. Nun: moo. do: . co?
3:: 93388138.
>mn 3:me NAN uoNo 32 ES «moo .NN: 33A
«Mao—8.80:» NAN 3AM «A _o NoA 3mm. oNNo mod
.38— 888 :1an 535 NAN Nu mo o.N ob No So
9.0mm maom 31:83 WES NAN N._ Pm o.N ob N.o m.—
Fa<o8mo 32o NAN o.oo o.Nm ouN o.oo on: o.oo
50an 8 88— 88a NAN o.oN o.oo o.oo PS 93 o.oo
€355 818. 8 88. 888 NAN o.oN o.Nu -o. _N .93 o.: oNo
Hgmmza 8882.9:— 886 NAN vo oNo o.NA 93 93 o.mN
008268 033.388 ESSEEE
08238 8:00:88: NC Gm NAN o.o o8 N; ASN
9:288 8.25.8 832 any; N; as was a: at. 3a :53
En 9.5833: Eng
9582 Gomsanamluaa
.38— 888 :1an WES .3 SN .A.A ob om 3a 38
rn<o~wmo mu o.oo o. _A o.oo o.oo o.oo ohm
~8an 9:an Wow/S on No Goo 3o woo m3 NAAo
~8an none—moo: mu o.oo oNo .o. .A o.oo oNo o.o—
_:no_:o\_.o8_ 888 .3 o.oo o.oo o.oA o.oo o.om o.oo
Hgmmza 8882.08— mmmoa H3 o.A_ o. .N o.wA Pt 93 o.oo
main; oak Shiatsu”. $5.? mawznﬁazoz
wear 8:5. 1.3 NS foo o. 8 o.oo o.oo o.oo 93

5383.5: ESQ ax: NB who INA o.oo o.oo foo o.oo

39

Hue—o r— 2.53:3. mazwaam 8.. $8.88.— 88 Don—28:3

Moik82ma hon: 083882.28...
row: 2:95. 373
78813. 88.3
89.8:
358.88 @3838
408— 2.882 0.. 8:88
.88. 8:82 8.. 8.8 2352M...
,8...— ::.:c2 2. c.5888. 8:82....
.88— m::8 :88 3. 8.8 2852......
00:08:28.8: cm $5988 @8388:
.88— ::.:c2 on, 885: 8.8 22.52%.
898 .815 0:3. 885: 88 22858848885
88% 815 m .858 885: 88 2352488.:
388:8 cm 885: .8: 2852Amx888vo
2:872 0A. 885: 8.8 28588 32

88.3... 003.838: QR. N «on... 8&3
Z885. 8358.8? 518—88 Ao. 3
38:82:15 AN. 8
#258; 838.. 08 :8 E8283. AAV
4.8% Go
”Q: 288 A3
m2<88m Q. 3

$588 02.8: 85.88..

082:. 8888 8388;255 838—
.88888 8: 8:53.88
~50:st
08. 82.88888. 86888.8
092
3.88:2 83.8.3
088. n5.
052

3;
Zoo
3N0

.oom
_oom
_oom
_oNu
_oNo
oooo
~oom
_oom
Noon
Noon

_oom
_ooo
.oom
_ooo
~oom
Aoom

Sow
goo
goo
.oom
.oom

.ooo
.ooo

_._ F28 2852... 88.58 88385 8 £8.88 on woo}

uoN
NNoA
o.NN

on
PA
u.—
N95
.353
_.A
o.No
o.o—
o.uA
ohm

o.oo
ou—
o.wN
o.oo
o.o—
o.o—

o.oo
o.ou
o.oA
o. 3
95

on:
o.No

meo
53
92

MN
m...
A.—
No.N
No.oNu
No
9.5
o.NN
ouo
o.Ao

o.Nm
ouo
o.Aq
o.Nu
o.oo
o.:

95
o.N—
o. 3
o.oo
o.NN

PA»
93

Ao
5.3.

w
_

o
_o.N
oAAmo
o
o
o
o.oo
o.oo

_oo
29..

u

N

N
No.o

NmoA.m

_

o

o
o.oo
o.oo

N3
NmAw

o

A

u
No.o

moooo

N

o

o
_.oo
93

moo
Ammo

_A

moo
moooo

_.oo
_.oo

40

an!» _.N 9.5324. mangaem 8.. manna-2— 8»: Una—m w! Oran—8....

.25 82a 9883 8855 £3.38 8.. :3 8:6.» 8». .8No $58.88.. .25 gnaw 8.888525 NAN 38m «88 .37. .oeN 8 magnacﬂ
No3. 99:38 8n an: 2:. 8n 0.50.8. 0:888:38 m3 8.8.22. 2 9a :3. .96. Ba 0.508. .98.. 38836.3 $2.38 .9 com.
orﬁwﬁoamaa 86 8.85.2. a. So an». .95.. ..wno-8..o:== 8:. ...u0m?38:== 98 amazon 882:5 8 2.592 :5 $9.5 <81. 8.. 8a
.8: 982.3 .89 0.588. mam—E8: .m 88.52. "883.5 8 =5 .Cﬁd 2:. 8 ran .33..

 

 

 

H88. .52.— 0.30.8.
Ham. 8..
2:3. 05.2.88
03. 98-88. $8-883: $87883. 7.8.: 3-8.5..
2:88:
NASA
3.3: 0.338812%... «21.33. 85:25. AZnoS AZHNd AzHAc.
>ma 33a. NAN 3833.. 8333:. 9mm: .38..
MEEOVBQA NAN “33.33. .33.: .3. RAE—Nu: «3N: .muo 9Am 3 9N3“;
H08. 886 35.85 .035 NAN N833. A.N_A.A. A333. 9.: 92 9 EN. 9ANco
03mm 8.8 91:83 535 NAN 9.3.x. A83. .. A330. 933.... 9. Go 93Ac
Egan—mo Baa NAN 933.8. 933.3. 9.33.3. 933. 3. 983 983
.soan 8 88. 888 NAN 983.8. 983 .8. 983.3. 923.3. 9. .8 9NA3
$9.35 26.8. 8 88. 388 NAN 9oN3.Nu. -983. .3 5.83.3. .933. .N. 983 933
.3553 Emansgw. mmmﬁm NAN 9Ao3.No. 9A.3.NN. 933. 3. 9AN3.NA. 93NN 9aNN.
95538 6083538 88:83.33
0552,88 8:835:83 N .w .uthAb. Aux—9A2 9.2.93. 933.3. 93on 9mqu

Cannon—no 62208 838. imam

25 052%... 85... N; $6.3m. 33.53.23.32.8.8.8.8. 988 88..

98mg. 0:388338

\ZMZEF ox 28805 .2 v .2 v .2 .

#23888. 3....83 WW2. 3 3.2—AA. 3.2—N... wN.2—N.N. 32.9.. 983 9cm“...
r2038 3 9003.3. 92.3.3. 9333A. 933.3. 983 . 983

41

.320 r» min—Esq mains»... n2. mus—$33.. r9:— Uouw aw. nuance—£333

:8an 3.525 535 3

.3850 8212.0: “3

.snoBQHOS. 888 .3

4.9.5:...“ 8.5.2529. $88 .3
was»... 93% 32:583.... ESQ walEﬁaaoz

mgr 3:5. .9... Nw.

538:0: 3:3. .93.. 8.

@3388“. hon: 039323.25.
\ZHZE... ex .23....

how: 989.... 3.5 3.x
7.9.5.3. 38a. Eco
mooﬁon. 3%.
$33.83 $2.256
HOS. 2.3.x: 0.. .0302 Son
H08. 3:562. 0.. .35 928503.. .25
408. 256$ 0.. 25.0.25. Wanna... .2...
H03. mama so... 3. .25 gmaa. 5N.
03833.8: 0.. $35.88 gnawing: 2.8
H99. 2.332 c». 385: .02. gunman... .2...

.585 33.5 on? 385: .09.. magmaqwiacaavo .mom
5.5m .535 m 35.0 35m: .35. Bugging Sow
188:8 om moan: .02. magmaszEESBE 3cm

2:562 0.. 35m: .25 935mg... 33

8.9. 9:562 0., .35 923mg .93

32:8.
93.9%.
93.98.
92.95.

93.9....

93.. Z...

.333...

83.2.3.

98.9...

3.3.

9......
9......
3.295.
.89.
.933.
.39..
98.9....
9. :93.
93.98.

93.98.

8.2.3...

93.99..
93.93.
92.93.

.23.
93.93.
93.93.

9.qu...

53.3..
32.33.
93.9....

929..
2.93
9%....
N923...
333
.38...
929:
98.98.
93.98.
93.93.

9598.

63.53.

98.93.
93.93.
93.93.

e4nuq.
93.93.
93.9%.

AZHm...

3.3....
.8238.
93.93.

92%.
2......
92%.
3.3.9...
9.3...
.23....
93.9..
93.93.
93.9...
93.9....

93.93.

“3.5338

93.98.
93.99..
9.5.93.

QLHoq.

93.93.

93.98.

2"”..va

9.53.
33.39..
93.9...

929...
3.9..
93.93.

93.2.2

.39..
.88....
$.92

99:93.

93.98.

93.9....

98.93.

c. :8
9.3 .u
Pascm
o.w~..~

9.33
9.33

o.oooc
988
Page

Pacoo
o.oooo
Peace
9.58

9.82.

98.:
9.83
938
cow 3

Puma

c. _ mm.
9.3:
:52.
93.3

92.8
988

o.ooco
o.oooo
Pose

o.oooo
o.oo . N
oboco
o.oocc

o.ocoo

Poco—
9.2.3
933
Puw. _

o.NAwN

 

.. .62. 935mm; 30252. 383:5 8 $2592. an woo}.
.3 Io ”2o 95.238

42

5.20 —.N 9.53:3. weigh... n2. mix—3:2. Pom: Una—u aw. DES—3: Ana—.3

 

 

 

HS». HEN Qing—
Hom» ow
2:3. 05.2280
03. 9.6.88— ??Rmoad womznmozs 78$ GENE—av.
($588:
N-8m.
3.2: ﬁrawnﬁmluaa «ZHNSNS. o\\m§a.\ AZHEV AZHNNV Quad
>mo 3»me NAN .3333: 32333.: 343333 .33333 933 9m Gm
man—835:" NAN “333.22 mun—333 33383 “N. 2533 9.8.2 98%
H08— 2..QO 32:25 gas NAN N293 NENE N. :NS N29: 9 :3 9 :8
08% 3:8 32:25 awév NAN NLUE N.:N.u_ 3.2.“: N532 98.5 923
553% 3:0 NAN 933.8. 932:3 983—3 92:93. 988 988
5830 8 88. 888 NAN 9on98_ 93:15 93—99: 98:53 93:3 933
€2.55 83:: 8 8&— 838 N3 98:53 9399:: 9929—2 98—95. 9N_Nu 9N_ 5
Hun—min 8382.2”: 888 NAN aha—9N3 933N: 9u:9NNH 9AN~9 .3 9ch— 98 5
0°38an romwzaznm «353%?»
0550876 8:833:03 N5 3035.3 5.3292: 2.32008 5.32463 9am“: 92$
9:298 3.24: 833. 53... N; 323$ 3.538; 33:8 : 53.2.8. 98 a. 988
35 9,5296 :mzm
9:802 Grenaamluaa
\2u253. ox 23mg?» «2|qu 2| _3 AZINd _
H08— Emoa 91:82.. 535 3 5.23.3 HNGS Mmaﬂ 9.39: 9N. _N 938
Enigma H3 938. T: 9.53. _ & 938.3. 9029 _S 988 983
58:5 @588 ads .3 3252: mogumﬁ waif: ANN—33 938 9Nuum
5093 8:338: 3 98:53 98:5 : 9 389: 99:92: 9w EN 933
58:52.08. 388 3 93—99: 933.c2 9:398. 9:298. 93% 9 _ m3
Hgmmza mumﬂwaS. 3.83 mg 923. _N_ 9329—3 92:9 _ : 9529 SH 98%.. 9%3

43

5.20 r» 9.532.! meaning. 3.. 35:09.2— 5».— Unuw aw. Oran—3.... Ana—.3

main: Si 333205.. Mei? wawanﬁnzez
mgr ESQ ax; Nu—
gmazaos ESQ ax; H:
@zaqna‘mu hon: ﬁxaxmamlmna
\ZHZSF ox N25,:

row: 2:95» 373 a;
73256. 833 Sea
mmnﬁoq 33
$58.33 @3256
H08. 2552. c». Emma; 33
Hog 3:352 ow Ema ”5852.4... Sow
Hon: sea—won a». @939nt 5.502.... 33
Hog: £58 ~53 3. .25 ggmoa. SB
00582330: 0.. $5988 Eon—753:: 38
9.2». =cho~ a». 88mm: 53 Ema—d... Sow

F83 3955 on; magma .83 magmnaaazaawv 33
P83. #935 m ﬂaw—a ”dam: Ema magma—.4358 Son
3838 on 33mm: 53 magmozmeF—BBE 33
29552 cm 38mm: 53 magma; 96.,

8:: 2559 cm Ema 23:QO 33

329::
93: S:

8282:
89:223.
98.92.

9292

9:9:
9:92
8.28.2
59m
.833
929.:
529.5.
9 _ :99:
93.98.

8293.

2&8 2&3
98.22 92.3:
93.93: 93:3.
Eugs 215
3232 363
$323.32 .3533:
93.9.5 98.93
929: 929.:
9292 929:
9:92 9293
N923: 8.592
38.“ ~89.
3:92 ESE
22.: 92.2
95.9w: 98.98.
98.9w»: 93.98.
93.93; 93.98.
9583 9292.

A215
523:
3298.

AZHNKV
umwmmwos
Sag—32
c. Snob:

33.3

9292
9:93
3593
an:
B392
5:2
93.93.
9_ :8 :
93.93.

98.9w:

o.Nﬁw
Pc—No

ch _ mg
933
P88

Poco—

o.ooco
cum 5
988

P88

0.33
985
93$
93:

983

chwom
o.owqw

o.oooo
o.Nuuu
cbocc

988

988
cud:
obooc

o.oooo

93.3
PS;
93%
933

93.3

 

_. Puma magma; 33:an ”noon—Sm 8 map—Ema w. woo;
.3. to u 20 E2238

44

Hue—o PN man-Esq 9.338 3.. mix—5:2— Fon: 62:... av. Oran—3r Ana—.3

 

 

 

do.»— 22:33:.
Homn on
25:. 92.238
Ocm. 95-86— 39833 $878323 78% 9&3ch
(5.098:
NAnm»
3.2: Quasﬁmluaa Q<H>€3 £5333.» ﬁZH—u: AZuNav Q4"— 3v
>mn€mwmv NAN .BNNEHX: 33333 mNmATSwﬂ $33.33 o.Naoo 93.:
m32333. NAN Noam—.23. 333333323ud NNSGBS 982 983
408.888. 3533 Ewév NAN N293 NdmﬁNd ”£293 Nuwﬁba. 93NN 9waN
08mm «38 :1an aways NAN 9:93 .62wN3 r839: PEULB 933 9&3
ra<o3mo 83o NAN 9am—9N3 933.3. 9929:: 989.01: 953 98%
SSSoSSEmmRG NAN 98:33 3529—3 988.38 9299—: 9:3 9qu
225% 838. 8 SE. 888 NAN 9oN_9Nu_ 9o29N£ 9a.: 93H 9829: 933 953
Hgmmzo amonmqaa ammonm NAN 93:53 92553 93_ 9N3 93:53 9.33 9umNe
QQSQSR Qo<m§a=nm «ozzmaih»
095236 8:83.38: N: 3.23.3 NobomNobE—qbctuwﬂ3.3.362 93.3 938
c5238 gémgsaéamza N: 3.233 :aNESZobEwE3.3.3.53 98: 983
E:— 95.23% 2E5
QﬁmvoN 05368138
\2u253. ex 23ng
H9»— 8m2m $1503 $915 mu SNH 3.:
P223." 3 9.68. 7:
58:5 92:25 525 3 ONE—mos
5830 83230: 3 98:53
FocBnDdE 388 mu 933.8.
H8535 8332.2»— 386 “S 9.23. _N_

45

2.96.0 V» min—5:2. @9238 2:. 9.52252. 5.»: Una—m 5. 06362.: 32.3

main: can 92:585.. 61:5,? 5:222:38:
mgr 2.ch ﬁx: N:
53832. 2.ch ﬁx; 8.

Mvzumnﬁma has: 062322828
\2u253. ox 898»

be»: 2:95. 32: 55
3255 33.3 Zoo
mango: Ema
M33288 642.2228
Hos: 3:362 2. .2598 Sow
3.2»— =c862 om Ema 222523.. 38
3.9»— :ca62 c». 33362: .2522 Row
3108— 320 626 6v. Ema gunman... 33
09.83228: 3. $56388 222.56%: Emu
3.22 2.362 cm 385: as: 23.52%. :8:

625m 635m on;~ 2.23m: Ema gmoaAacaawv :8:
625m 692.5 m ﬂaw—a 335: Fan ”ﬁgmnqlaﬁsa 3::
389.8 cm 2635: 5mm magmoAmvacaBE Sow
2:562 cm 2.29%.. an: 2.2.m2m 32
88— 25.62 cm :25 magma—w.

_. 63: 84253 89:53 moooamsm 8 9252: an :82
1. mo ” 2o 93.2280

38

2.233
8.3: 5.:

82823
89:38:

98:1:

8.29:
8.28.:
8.5.3
828.:
888.6
288.3
.23:
8.88.8:
S :33
328.23

8.88.3

aux:
3:3:

3:38:

AZHMw:
A—umwmoS

882883

983th

93...:

8.83.3
2:33
8.28.3
88.8
.88.:
2:3
888...:
3253
8.38.23

828...:

AZ" .2 c
328.:
923...:

A2": 5
8.28.:

88:88:

988...:

33.3
8.28.:
82...:

8.28.3

8.88
.88.:
5:3
8.8.8.3.

328...:

328.23

923...:

215
8.28.8:
93:33
233:
.3288.

88:23

983...:

92.23
~28:
8.28.3

8.28.:

88.8
.88.:
2:3

888...:

8.83.8:

8.3.33

988...:

92::
9:3:

95%
Pam:

9:3:

Page
a. _ N3
933
98$

93un

c. _ mom
oboe—
o.oo:—
933

9.80:

o.oo:—
936:

o.ocoo
983

©2396

o. _ Mme
938
9366
90:3

obwwm

ebu— _
o.ooo~
o.oo:—
933

9:6?“

46

266—0 :6 06:060.: .8: 382020: 60:: @2208..."

46:: 86.0 20:28 0003088 02822 «88 «02088:: 88:65 :60 $82020: 8:: 2.8220 :0 06:060— 936288 :60 :0:0::0::
«.2868: 20 :60 :2:8::6_-::8068:: 8:0: :12 8:: .4620: 8 :80. > 8: :::0_ n. :60 8:2 :620 0:. :60 8:: 60:: 6% :60 8::
222523 8 :80. w 8: :::0_ U. :60 20202.88: 20 2.6820: 6v~ 2:85 8:2 3:22 :063. 06:060— 8 : ::88v. :62 :28:
0: :60 .880 0:- 80 a. :60 :5: 8 2.6620: 8 : 06:060_. 8: :20 0:62:58. ,22: 06:060_ 8 : ::88v. :62 860: 0: :60 .880 0:. 0:0
:- :60 :8: mm 2.3620: 8 :8: :.<0 06:0606 8: :20 0:602:80. .22» 06:06:: 6: : ::88v. :62 :28: 0: :60 880 0:. 0:0 :. :60 :5:
6: 2.6620: :0 : :0:-8: :.<0 06:06:: 8: :20 0:62:80. 06:060— ::.:..:::8: 6.. 80850: :002:8m 8 :60 W20: :2: $80
008858: GAE-Q 8: 600 666$. ”08:: :m : ::88v. :62 :28: 0: :60 .880 0:- 0:0 3. :60 22: :20 0:, :60 8:: 6: 2:2 $3.
8: :20 0:602:80. 60:28:06: :0 :60 8:288 0:- :60 :5: 8::— :mm0:m 80820: 8 8058:: 0:. 886 820:: £08 60<2:m0 mm
:60 .28 0:. 8::_ 6:62:22 8 8::— :mm0:m. 800802-92 800:: :m :60 8:8 0:. :0: 80080 8 8::— :mm0:m. 60m::m0v mm :60 8:228: 0:.
:60 ::862 0:. :3: 2:00 :60 800::8: 0m :60 :8. H8268 28082.08: 8.0.0:: 6: :60 .28 0:. 8:868 :22: 8 8::— :mm0:m. €255
8:6:— mm :60 6:420:00 60300: 0:208 :82: :8: 0:208 6:65:09 60m:0:: 82:13: :0 :60 88688 0:- :60 82:10. 0:. :60 8::
8 :38. 6520:: :8083 :m :60 82588 0:. :60 8:: 2:0 8 Cm 8:22 m0020: mm : ::88v~ :62 860: 0: :60 .880 0:. 0:0 :- :60
8:: :0 00020:. :8: :20 0:62:80. 2:862: 8 :2086002 :0:0:0 86:2 28:2: 288 08220: 2 :60 :8: _0<0_. 3.... 1.. 8: ..
8:820 22:28:? 058308: 2 :60 .98. 6:8. 8: 5.x. _0<0_.

 

 

 

:80: >. 22:8:2: :::0_ w: 60:: 620
E :8 6: 3 :3 :3 3 a:
06:060— -: 6.3... 666.3... #6:. .m 6:966... 6.6.7... .62 $66... 8.9::
63:8 6:6.qu 63.3: 666. :3 6.68 6.3V 6.8V 6.3V
06:060. .. W088 $566.... 6563...: 62:61:... «66.31:. 93...: 6.3: 666...... 6.66
666:8 662:8 62:3 6.063 6.3: 6.48 6.68 6. :8
60288:: .5822... $3.68.: LP: .63.. L66: 6.6.:
60.68 6603 66.0.3 663 6.63 6.63
60<2:m0 68631. 65686.... 66:6». -93 $6: -206...
69.63 666.63 68.6.: 6.68 6.63 6.5
6.0080268— :mm0:m 6.86.63... -6. _ 666:1... 6.66666... 6.6: :6: o. 3
669—3 28.663 63.3: 26.6.: 66:3 66:8
60280: 36.6.. 63

69—3 6.3V

47

8.6.0 66 06:060.: 8: 986.020: 60:: 9808—6808..

...::m.6.0 :::0::\.- 08. :::0:: .366 L. . 6
6.3.6:. 6.3.
€268m 0:88:08. :::0:: :3. .6 L. .6
66.0.6... 666.
6028:: 82:10.. -6666 -66..
A5666. 2.5.
6028:: :80:::. 6:63.... 666......
66.3. 6.66.
m0020: -u . .66 6.2.

. 6.0.3. :83
<02. 8888: <0: <0: <0: <0: <0: <0: <0: <0:
88:5. 8880: <0: <0: <0: <0: <0: <0: <0: <0:
60:: :E.:0:0 888.0: 20 20 20 <0: 20 20 20 <0:
06:02:68: .86 :6: :6: o... .63 :6: :6: :3
”-2820: 6... 6.6 6.6 666 6.. . 6. .N 6.6 666

 

48

0.0.0.0 Pu 0.80.8.0 80 3.00.880 r08 mag—0.5.0 .000...

8.0.0 80.0 00.0030 00030.08 0008800 3.08 009000.000 0080.8 80 080.0880 .08 05.088 8 0.8000. 03.8000. 8.00 0000000...
8080.00 80 80 E03008.-:.~00.88 800x .0. .08 0.8.00 8 380. > 80 .080. O. 80 88. 0.800 0.. 80 .08 70.0 0% 80 .000
2.808%. 8 .080. .w 80 .080. U. :0 080000.000 000 00:80.00 3. 2.0.85. .000. 098000 .Ormv. 0.80.00. .0 0 0.88% :8. 8.80
0: 0.0 8.00 0.. 0:0 ... 80 8.8 .0 8......880 8 0 0.8000.- 80 ~08 0802.00. .20... 0.80.00. .0 0 00880. :8. 8.000 0: 80 8.00 0.. 000
..- 80 m8 .0 03.8.00 8 m 8.0 .70 0.8000.- 80 N000 0803.00. .20...»- 08000. .0 m 0088v. :8. 8.80 0: 80 8.00 0.- 000 .m 80 .-=8
.0 03.800 8 0 000-8.. .70 08000.. 000 008 0803.00. 0.8000. 03.88: .0 808.300 0000008m 8 0.0 .8000 10.0 .0800
0088.00.00 aha-HQ 80 .00 3003. W088 .0 0 0:88v~ 08. 8.80 0: 0.0 8.00 0.. 000 ..- 80 08: 00.0 0.. 80 .000 .0 0.0.0.. .00...
80 N03 0802.00. P880030. .0 80 .0m8..88 0.. 80 38 88. 0000.0 80008.00 8 80:00:00 0.- moc... .0080: €00. P0<08m0 .0
80 800 0m 88. .88....00 8 88. 0000.0. 80080: 08. 0000.0 .0 80 800 0.. :0. 80080 8 88. 0000.0. r0m8m0. .0 80 880.88 0..
80 8.800., 0.. 08.0 0800 80 8003.0: 0.. 80 m8. H8820 0000.02.08. 0000.0 .0 80 8.8 0.. 8.5.0.0 0000.0 8 88. 0000.0. €00.08
00.0.8. .0 80 05.0.0000 00.200: 9508 0000.0 80 00:08 .88....00. r8208 8080.60 .0 80 880.88 0.. 80 8080.0. 0.. 80 .08
8 08.0. .nom..00= 08008. .0 80 88.-.88 0.. 80 .000 0.00 8 Cm 00.800. m00§00 .0 m 0:805~ 08. 8.80 00 80 8.00 0.. 000 ... 90
.08 .0 0000000. 80 008 0803.00. 2080000 8 0000080000 0000.0 8.000. 080008 02000 0.00.0000 0. 90 8.8 .0<0.. 23. t... 80 ..
80.080 0800000.;~ 0.8.308. m. 80 .o\... 0.8. 80 8.x. .0<0..

 

.080. 0” 201-800... LE-mbﬁnll

 

 

.8 L8. 2.. .8. :0. .2. :0. :3
82.388. -203- -033. 0.0.8 -083. 0.3:. -93: -90.: -98:
.0030. .0300. 0.0.5 0.0.5 3.00. .000. .008 0.00.

do: 983. .. .633. 93.8... 22.2: 30.00.: 30.:- 000: 0.3.. 0.0... 0.3
.0095 000.8. 0:00. 000.00. 0.00. 3.3. 0.00. 0.03

:20 0.88. -0030 .0000. .088 .0200. -000 -000 -03 -000
000.00. .0020. 00.0: 000.03 0.5 .00: .000. 0.3.

82098318093. 08.00: 82-3.: 53.8.: 83.3.: 0.00: 0.3: 0.00: t:
08.3. 000.00 000.3. 03.0: .000. 0.03 .3: .03.

50.80%. 000.3... 08.00:... -200 -90.. -08 9%
30.0: 30.00 30.03 8.8.. 8.03 3.0:
022000 r880: r830. 08.: 0.3 -00. -30.:
300.3. 38.3. 300.00. 3.30 .30. 3.00

5830808. 802... 0.80.3... -0. .8... _ :. 0.93.00:- 0... 0.: 0.3
0:900. 003.00. .8009 00.00. 20.08 20.3.

49

Haw—o ru Oran—5.8 3:— 38:83; re...— mséancnn Ana—.3

roam—mow 59m— foo
39.88 2.2:
Hgmmzo 888832”: 8828 :83 -83...
$3.08 $.me
€3.35 BBS—DdE 880$ m¢m.~u -NQN
3.3.x: 0.3V
romcog =58an .mfmq -o. 3
:59: 2.2:
romcog 2595: -amc. 33.... -933...
89m 5 8&3
moncqna -NPum Pea
Awawud ﬁbav
<9: 9558 <8 <8 <8 <8 <8 <8 <8 <8
=3:an 955:8 <8 <8 <8 <8 <8 <8 <8 <8
Fog 9558 95:58 20 2o 2o <8 2o 20 20 <8
0:829:08 SB 23 23 o: SB 3.» 23 c:
FEE—8a PS 93 93 93 PE 95 93 ohm

 

50

2.8.0 .3 0.80.8.8 0:0 3.00.0080 r08 mag—0.5.0

.28 8.0.0 «0.00:8 00030.08 8:888 .._.08 808088.08 8.8.8 80 880.080: .00: 80.00880 8 0.80.00. 03:88:. 2....0 00:08.08
88.0.0.8 80 80 8.802 0.. .000 8.88:8 .: 00:0. > 0:0 .880. 0. 0:0 80 88:0. 0.. 08.0.08. .8008 .: 00:0. w 0:: .880. D. 2.00
80988.08 2.0 08:80.00 3. 08.85‘ .08. 8008.08 .O..m.. 0:00.00. .8 0 0:88.. 88 8.808 0: 80 8.00 0.. 0:0 ... 80 m8 .8 8......800 8
8 030000.. 0:0 N08 0803.80. 2.02 0.80.00. .8 0 0:88.. :8. 8.88 0: 80 8.00 0.. 0:0 ... 80 m8 .8 0......080 8 w 8.. :8 0.508..
80 N08 0803.80. .205 0:00.00. .8 0 00880 80. 8.88 0: 80 8.00 0.. 0:0 ... 80 m8 .8 8......800 8 w :0:-80 :8 0.808.. 80
N08 0803.80. 0.80.00. 03.820: .8 808.300 00008.5 8 80 .8800 mm: 2.300 0088.888: 203.0. .80 P00 Eco... W088 .8 0
0:88.. 80. 8.88 0: 80 8.00 0.. 0:0 .0 80 88: 00.0 0.. 80 .00: .8 8.0. .00... 0:0 ~08 08032.80. P080808. .8 80 888.88 0.. 80
..=.8 88. 880.8 80850: .: 80:80:08 08 M08: .8800: (<0:. ..0<08m0 .8 80 8:0 0.. 88. .83....08 8 88. 8808. 800802.08.
0880.8 .8 80 8.0 0.q :0. 80080 8 88. 880.8. r0m8m0. .8 80 .5388 0.. 80 88:8 0.. 00.8 8800 80 8008.0: 00 80 m8.
2.88.0.0 0880.82.08. 0880.8 .8 80 8:0 0.. 8.8.0.0 8808 8 88. 880.8. €08.38 000.8. .8 80 0.....08800 002.00: 950:. 0880.8 0::
0.50:. ..0.0.....08. romeo»: 888...... .8 80 .0mE...:8 0.. 90 80.8.0. 0.. 50 .00: .: 03.8. romeo»: 0808.. .8 80 .0m8..:8 0.. 80
.00: 8.80 .: Cm 00:088. moo—=00 .8 m 0088.. 80. 8.808 0: 80 8.00 0.. 0:0 ... 80 .08 .8 800080... 0:: N08 0803.80. 2080088 .:
0080:8088 00:08 80:8. 88:00:. 0:08 0.08.0800 0. 80 .._8 .0<0.. 1:... .3... 0:0 ._ 80.080 8.8.8.8.... 8.8.3008 0. 80 .0... mo... 0::
5.x. .0<0..

 

00:0. >“ 208. 0.. F000 >Q8m088

00:0. w” 208. 0.. 03.0.08. r0808

 

 

:. .0. .8. 2. .8. .0. 2. .8.
0.8000. 908 9.0 90M .2...» 9:0 9 .0 9 .0 9 .N
.92.. .9... .900. .93. .93. .93. .9... .92..
0.80.8. _. x088 -98 9.... -900 -9... -9... -90.. -900 -9...
.908. .9... .900. .93. .90... .90.. .90.. 9.0.
rom.880.8. 9 3...... 9 .m: -93 98:... 9.. . t... 9...
8...... .900. .900. .90... .900. .90...
r0<08m0 ...u_.:._ .20.... PEI... 2.3 ......0 -9~..
.98... .9... 8.0. 2 .o.. 2 .00. 8.x...
800802.08. 0880.8 ..... 93 n... .2. ...m.. -..m~ 9....
2.8... 2.3. 28... 2.0. 2.... 2......
rom.0m0. -90.. -9N0
.9 . u. .90....

51

2.020 .3 0.80.8.8 8.. 0000.00.00 P08 0.3080280...

2.80.0.0 0880.82.08. 08808 T... 900
.0... .900.
€01.30 80.8.2.2». 0880.8 .0... 9 .0
.93. 2.0..
homes: 308...... 0.. . -0. .0
.0. . .. .9 .0.
r0020»: 80:8. .003... .002...
.0. .0. .0. .0.
000500 0.00.. _. -0. .N
.90.. .900.
<08 0:88.08 <08 <08 <08 <08 <08 <08 <08 <8
.8883. 0088.08 <8 <8 <08 <8 <8 <08 <08 <08
.80: 05.0080 083.08 20 20 20 <8 20 20 20 <8
008028.08 .000 .0? .MNA .00.. .000 .00.. .0...» .000
9-8000800 0.00 9.0 9.0 0.0.. 9.. 0. .0 0. . 0 0.00

 

52

0.050 —.a Oran—00.0 2:. 00.85.80 r00.— mnanawo A000».

.28 8.0.0 00008 00030.00. 88.0080 038 000008.08 00.0200 80 00.00.0300 .000 3.0088 8 0.8000. 03.8000. 0.00 00000000.
«8.8.8 80 80 0:00.000 0.. .000 830008 .0 00:0. > 000 0000. 0. 0:0 80 00800.. 0.. 0030.003 .0808 .0 .0000. .w 000 .0000. D. 0.50
«00.08.08 80 8030.00 3. 08.05 .08. 3:88 Aormv 0.8000. .0 a 00880. 80. 8.80 0: 90 8.00 0.. 0:0 .0 90 ma: .0 03.880 8
0 000000.. 000 N03 0803100. .20: 0.5000. .0 m 00:56. 80" 8.80 0: 80 8.00 0.. 000 .0 80 .050 .m 03.8.00 8 m 8.0 020 070000..
000 N08 0803.00. .508 0.3000. .0 0 008300 80" 8.88 00 80 5.00 00 000 .0 80 ma: .0 03.8.00 8 0 000-80 30 050000.. 28
. N03 080280. 0.5000. 03.8000 .0 800.500 00008.00 8 80 .0080 0.0.0 .0800 0058.88: 20.0.00. 000 .00 8000.. ”000:: .m 0
95:30 0:: 8x8 0: 80 8.00 0.. 000 z. 80 08: 00.0 0.. 80 .00: 8 0.00.. .000. 000 N03 0803.00. r00A8008. .m 50 808.58 0». 80
0:0 88. 8008 308500 8 908008 00 00:... .0000»: £00. P050000 .0 90 800 00 88. ..00...08 8 88. 8008. 500802.08.
8008 .m 80 3:0 0.. :0» 80030 8 88. 800%.. r0088. 8 80 808.05. 0.. 90 008000 00 08.0 0.000 90 800.000: 0.. 80 8.3.
0.80.0.0 80082.08. 8008 .0 80 800 0... 8008.0 8008 8 88. 8008. £08.00 00.0.8. .0 80 0.0008000 007200: 00:00. 8008 0:0
95.03 .82....8. r0020»: 3080.00. .0 80 808.05. 0». 80 80816. 0.. 80 .00: .0 0.88. r0028: 05000.. .0 80 808.88 0.. 80
.000 0.8 .0 C0 00:08. 000500 .8 0 95:5. 80. 8.80 0: 80 <0.00 0.. 0:0 .0. 80 .00: .8 0005.00. 8:. N08 090280. 208008 .0
0800888 00008 808" 835080 0208 0.8.0000 0. 80 0:0 .0<0.. .3... .1. . 000 _._ 80.008 088000.? 0.02.0000. 8 80 .00. 000. 000
.000 .0<0..

 

 

 

.0000. O” 203. 0.. P000 ggmﬂm .0000. U” 203. 00 0800.02: r0808

.0. :0. c .. CN. :0. .3. c0. :0.

.20: 0.8000. 70.0 .00 .00 .00.... 0.00 0.0. 0.00 0.00
8.00. 8.00. 8.00. 8.00. 8.00. 8.00. 8.00. 8.0..

.20... 0:00.00. ._ W083: -0.00 -0.00 -0.00 -0.00 0.00 .03... -..~0_.. -0.00
8.00. 8.00. 8.00. 8.00. 8.00. 8.00. 8.00. 8.00.

H.000. 000000. 0.00 0.00 0.00 0.00 -0.00 -0.00 -0.00 -0.00
8.03 8.00. 8.0 .. 8.00. 8.00. 8.00. 8.03 8.0 ..

ﬁﬂw 0.80.00. .. ”0002.0 . -0.00 -0.00 -000 -00. 0.00 0.0~ 0.0 0.0.
8.00. 8.0.. 8.08 8.00. 8.00. 8.0... 8.00. 8.00.

r00A8m08. 0.0... 0.3 -000 0.8.2.... 0.0.0.2.... 0.00
8.00. 8.00. 8.00. 8.00. 8.00. 8.00.

P080000 .001... .003... .003... -..N0 .000 -0.00
8.00. 8.0.. 8.00. 8.00. 2.00. 8.00.

500502.08. 8008 0.0M 0.00 0.00... .. -.00 - . .0. -00.

0.00. 0.3. :00. :03 :00. 0.3.

53

Ana—o rs Oran—5.8 2:— 38:33.— rona man—.256 A353

romAmmmv -ch -oba
. 8.3V 8.2:
HgmmEo 88282.2”: 888 Fact. 93
8.3; 8.3V
£0195 8385.08— 88on o.oo .98
3.5 Cbov
romeo»: 385:3 P: -9;
8. _ c e. E
romeo»: 2:055 #33:. _.w__.:.:._
8. .3 8. :3
moo—=3 Pact. .93
8.3V 8.3
<om~ 955:8 <8 <8 <8 <8 <8 <8 <8 <8
5955‘ 955:8 <8 <8 <8 <8 <8 <8 <8 <8
F25 9588 955:8 20 20 2o <8 20 20 20 <8
06829.88 33 UNA UNA 53 Son UNA EMA $3
”-8958“. o.oo o. .o 0.: 93 c. .N o. _ u o. 3 93

54

 

.768 —.u 06:68.8 8.. v.23": F888

H68 868 889.8 883888 83888 3.08 now—.8888 888m 888: 8:88. 88 8 ”5885 8888888 888 8 06860. 83:88:.
noon-88:6 8 «38888 88 838888 6% 6868 .28 8.8888 8:868 8 88888: A: 8 A3 8 w 888v. 88 :88 o: :8 <88 cm 28 :-
88 8.8 8:888 88 88mm: 688. 88 268288 . d8 88:88 8:868 8 8m8888 Ad 8 28 mm m 88.5. 88 8.88 o: :8 8:8 8.
28 z. :8 88 24ng mm m 8888 28 88mm: 68.». 83 088288. 06860. 8 m 888% 88 898 o: :8 <88 ow 08 z. :8 8.8 8 83—888
8 a 86862. 88 83 088388. .28— o686o_ mm m 88.3. 88 8.88 o: :8 <88 ow 28 x. :8 m8 8 83—888 8 m 88 98 8860—. 88 83
26838. 4.8% 8860— mm m 888% 88 888 o: :8 <88 on. 28 3. :8 m8 8 83.888 8 w 88-8.8 38 06862. 88 83 26838.
06868. 83:88 8 88853 mono—.88 8 :8 808» mm: H88 008888: GAE-Q 88 ran 383. wad—.8 mm m 8882 88 898 o: :8
<88 on, 28 2. :8 88: 88 cm :8 88 mm 88.. $3. 88 83 08838. romAmMQOv mm 88 .5888 on 88 8.8 88. 888 88853 8
88888 c.- mocz. K088 «<0:. 6988mm 8 =8 88 on 88_ 86:88 8 88— 888. 8888288— 888 mm :8 .88 cm 8" 8888 8 88.
888. P8888 mm :8 88188 cm 88 8.868 cm 85 888 So 8888: on 88 m8. H8868 88882-88— 888 mm :8 88 cm 8:868
888 8 88. 8.88. £255 888. mm 88 82.8888 682.88 08.88 8.88 88 9588 86:88. romeo»: 8850; mm :8 88188 cm :8
888.5. 8». 88 88 8 83.8. romeo»: 898$ mm :8 8m8188 0.. c8 88 888 8 Cm 8:88. wooing mm a 888v. 88 888 o: :8 888 cm
28 a. 88 88 mm 8888. 88 88 268298. 2:86.88 8 888688 8808 86:8 88:88 2.38 38888 8 :8 38 8.8—. 2:... 2.. 88 _.
8988 88885 888888 8 En _..\o. we}. 88 Ex. 886

 

 

 

3:8. >n 886: 688 >=8m88 588— w” E58 monomma >28m8
E 3 E 3 9 3 3 3 A8 :8 c c :8
06.860— 982. -981... -93.: -93....- -9:.:.._. -933...
8&3 8&8 8&8 8&3 8&3 3&8
06860: .888 9mm 9% 9.2... 932. 93.... 93...
8&5 8&3 8&& Ba: 8va 8&3
.:«106862 -98.... -98.... -93....- -9.S!.:.. -93.... 9.3:...
8&3 8&3 8&3 8&d 8&3 8&3
do: 8860— .. “888 98.. 98... 9mm... 9mm... 9mm. 9%
8.3V 9.98 8.on 8&3 Bad A99:
.28& 8860— -932... -931... -933... -98....- -9wo_...:.. -983...
8&3 8&3 8&3 8&3 8&3 8&3
donw 8862... .888 93 93 9E 93...... 93...... 98......

8&3 8&3 8&3 8&3 3.ch 3&3

55

romﬁmmnav 93...... 93......
8.3.3 A933
r8888 -_&m...... Luci...
8&8 8.33
88882.88. 883 9.4.9.... 93....
2.83 2.83
romAmmov -98
A933
3.882.. 8882.03: 888 -9:
8&3
€3.98 0885-88— 888 5.:
933
boa—o8 88553
58:88 8883
man—.88
<82. 888.8 <8 <8
8885. 888.8 <8 <8
88 9588 88: 20 28
0882888 3 _ m 3 _ m

Haw—n rm 0.80.87 2:— 33.08: 5:888:33

93...... 93...... 93......
8.33 A933 A933
-93.... -_&w..:. -_.uw......
A933 8.33 3.23
Peat... NS... NS...
Cbmv 2.83 2.23

9.:
.99:
-9...
.98
.5.
9.3

93

.933

98

9.3

-99.

9.8

<8 <8 <8
<8 <8 <8
<8 20 20
58 G 3 G 3

98......
A933
-93....
A933
Poe...
2.83

93
8.33
93
A993
-98
8. _ 8
<8
<8
<8
58

93...... 93......

.983
-9.—
.9...
-9...
98

<8
<8
28
G 5

9.5
.5.
.9...
-9;
.95
9o.
98
9B
9.:

9.2..
A933

<8
Zo
3;

98-:
.983
-98
9.3
.98
99$

93
8.23
98
8.33
-98
8&3
<8
<8
<8
53

9;...-
8.8
-98
.9...
-93
8...:

<8
<8
20
G 3

9;:-
.98.
-98
9.3
.5.
9.3
98
.983
9B
.98

99..
A933

<8
<8
28
3 5

98-:
9.5
.92
9.:
-98
98

93
A933
98
8.33
A93
8&- 5
<8
<8
<8
Sew

 

56

.320 r: Dan—00:0 9:... gun—.5. 30:08:30:

Him 820 00:03 0:030:02 00:85:00 «8:: 00900083 «088% 80 $8820: 8:: 05.0800 8 0:50:0— mazmzo: g: 80 8:0 0:.
838:0 62:6 8 mango—:03 0:. 90 88083::m 33:. .90 :30::0:. $58—00 E0 50 :0nm::m:_-=:80:am: 8:0: 80 8:: 0:800
:: 550— > 0:: 3:0— 0. 0:: 80 88— 078.0 8:. 90 8:: :2: 3. E0 _0m: ggmozmv :: 550. w 0:: 350— U. .90 «800383 2.0
03308: 3. 9:55 .000: 3:200 AOrmv. 9.58:0 we? maﬁa. :0:o~0m 90 :0..00:8m0 09:3. 0:000 :5: 83030 gang _8_: m:
80 conesmsm 9.3. 03082 mm 0 83.3. :5: 8:80 o: 50 580 0:. 0:0 z. 80 an: mm 3.258: 8 m 030:2. 0:: 003 0803500.
.20: 0:008. mm 0 95:3. 92 8:80 o: 80 380 0:. 0:0 2. 80 3:: :m 03:20: 8 m 8: 9.0 0308—. 0:: N03 9:02:00. 10%
030.8— :m 0 95:3. :5. 8:80 o: 90 <~E0 8:. 0:0 2. E0 3:: :m 03:20: 8 m :o:..8: :3 030—8.. 0:: N03 0902300. 98060—
0358: :0 80350: 802.95 8 50 W20» mm: .380 003.3008: A39 0:: P00 883. ”08:: :m 0 95:3 :5: 8:00 o:
80 e680 0:. 0:0 z. 90 new: :20 0:. 90 80: :0 was. :03. ms: N03 9:03:00. P380005 :0 80 8808:: 8:. 90 an: 8:: $008
80850: 5 805080 c». moan: $30»: ’20:. r0<03m0 mm 90 2:8 0:. 88— :85:00 8 88. 8006. 500302.08. 8006 mm 90 :58
c». :0: 50030 8 88: 8008. romﬁma :0 50 88:95 0:. 80 3.560.. 0:. :8a 0::00 80 50028: on. 90 33:. H0585 80082.08.
8005 :0 80 8:0 8:. 85:20 8005 8 88— 00008. €355 03:8— :m 90 :500030 60300: 0.5.03 8008 9:: 0540:. 58:300.
romeo»: 30838 :0 90 8819:: 0:. 90 38:3. 0:. 80 8:: 8 :85. homes: 2:95: :0 90 Ewan—:3 0:. 80 8:: 080 :: Cm
8:08. m0050: mm 0 9.3.3. :5: 8:80 o: 80 51:0 0:. 0:0 z. 50 8:: :0 00050:. 0:: N08 2:012:00. 25:80; :: 3003:0000
:0:o~0 885: £983: 0033 080:2 m: 80 :2: _0<0_. 3...... .3... 0:: .. 5:808 08:335. amamoe: a 50 1x: u..\... 9:: 8.x. _0<0_.

 

3:1 >” 201.53:— vm:0_ w” r0»: 2.2.0

 

 

E E Q: 3 Gr 8:

0080020 95w mun—:3. No.8 3.:— Ehm o.om ob: ob:
CANS Cuba: 2 T23 8.: 3 8. .3 8. :8

DEERE mgr was? .._ W083: $03: -38: -88: 8.8 8.; -88.
8..on 89:8 8:88 8b: 8.~: 8.3:
95080. -wuo.:w 83.8 Lamb: ANN .93 1.8:...
88.93 835:3 Auooms 8.3: 8.3V 8:»:

950:0— . W082: r3932... 0.3.3:... 298:: «8:... o.oo Poo
88.0.: 83a: 8883 8.3V 3.83 8:3

088030 mgr map—5. _. 95060— .593... ’38:: 88.3.. 8.8 -c.- 8.8
29:3 2888 2 :88 8.~8 8. .8 8.~3

008030 mgr map—5. .. 0:50:8— .. ”08 3.8 SON 3.8m ohm 8.3 8.3

3.3.:

3:83

:33

8.3

3.8

8.0:

57

romﬁmmnav

ro<o8ma

58302.03— 8%;
r853

.3555 8828268— 8898

€3.35 haw—REESE 8806

romaog Bus—10¢
rowaom: 8:95:
magnum

<08. 95518
5985 955:8
r25 9558 955:8
OsmaZmaoam
”.3552.

an!» Fa 9.58:.” 55—» mug—:3 wnaamuuaoiaoarv

.393... -~_u.q~u: hwbq -78
AaONMV A338 Swat: 86$
rouwbm: r393... umubN Lam
33.93 336.: 38.7: . 333v
b.3959...» -mePSi. $8.398... «.3
$3.03 A5363 Awmmbmv A369
29%...
2353
38.3
Emu:
38.3
Gawbd
-313
23.qu
-3983...
Amomuv
.5ng
$3wa
<8 <8 <8 <8
<8 <8 <8 <8
20 . Zo <8 2o
emu cg o: 03
9.3 0.3 98 PS

Law:
8&8
Lu:
Amy—3

0.3
:POMV
:3
2.8V
-quw
3.ch
b.0—

3.on

<8
<8

33
o. 5

can.
8.3V
-e. 3..
3.3V

93
:N.m8

-98
2.5V
5.2:...
Sat
93
G. JV
<8

<8
o3
o.Nu

 

58

ﬁne—a —.a 505838 was.» mun-.5. 5.1882808033

58 8.08 «308 008308.: 88:88 $08 898888 8.88m 8o $588808 88 85888 8 038.00— »3880 28 50 ~08 0m
8888 3:8 8 8888828 0.- 8a 80:98.5 an? .28 028.82: 0880—8 08 :8 201-588-2838»: 889» 8a 88 8:88
8 $52 > 28 380— 0. 28 80 88_ £88 0% 8a 89: 88 3- 80 88 388:3 8 352 w 28 $51 U. .28 388888 80
830888 3- 0885 88" 8088 SEE. 00888 mgr 8ch 8098 :8 889.8% 3ch 8:98 :8. 838:." awn—Amy 88 8
:8 802038 30.8. 9880. mm m 888% 82 8:8 0: 80 88¢ 0»- 08 :- 80 9.8 mm mam—888 8 m 9880—. 28 ~80 080389
H82 £830— 8 8 95:3. 82 8x8 0: 80 880 0w 0.8 =- 38 33.. 8 03—888 8 w 83 «in 05860—- 28 83 0:82.80. H8-
0:800— : m 888% :8" 8:8 0: En 8:5 0»- 000 a. :5 ma: 8 ”3:23 8 w 003-86 98 2880.. 28 88 082.289 9880.
93—88: 8 89:58 88:85 8 :8 K08» mm: H280 00:00:88: GAE-0v 28 Fan 38$. W083: mm m 88:5~ 83 8x8 0:
80 880 0.- 000 =- 80 88: 88 0m :8 88 8 M52 33. 28 ~20 088280. romA8-88v mm :8 883.250 on "so 33. 8E 8808
808088 8 8088.8 0». M009 $08»: $00. 883% 8 En 8:0 0»- 88— 58518 8 88. 8808. 800802.08— 8808 mm 80 3:0
0.. :9 88:8 8 88. 888. romAmmav mm :8 8m88§ 0m En 00082. 0m 88 888 80 8838: 0»- Eo mas. H8588 8882.08_
888 8 80 2:8 0% 8588 8.88 8 88_ 888. $0195 SES— 8 :8 92.0808 8230: 00:2: 8808 28 952: 58:88.
romeo»: 8880.va 8 :8 8838.0 0w :5 83010- 0m :8 88 8 83-8. romcog 0:855 8 80 8m8§8 0m :8 80: 88a 8 Cm
8:08. moo—.88 mm 0 98:5‘ :8. 8.88 0: :5 $80 on 000 =- 28 8»: 8 88:88. 28 N20 0903280. 203028 8 380888
8:08 808. 88.888 3.03 0888“ m: :8 ma: _o<o_. .3...- E... 28 . 8988 8838385 885502: 2 8o 18. u..\o. 28 5.x. _0<0_.

 

350. O” T8388:— vBﬁ. U“ r28 2.80

 

 

3 as g :8 C c :8
82:02: 8% 35 ~98 .98. 2.3 99. 98 -92
:33 :88 0 83 8.. c e. _ c a. _e
08.38 we.» maﬁa - 83:: -09; -298 -23 -98 -9; -98
3:5 @980 6.83 A90: 8.0: e. 5
do: 385 -398. $8.: -398 -98. -98- 5. _ _-
83.80 03.80 30.930 0.08 0.0. c 3.08
in: 0832 - 03330 03.3... .59.;- 293 9.: 93 :0
$0.980 $8.30 883 3.3 3.0: 3.3
053% me; me? - do: 935 -3: a... S -09: -9 8 -9 a -98
$0.30 983 6.8.3 8.8 8.03 8.0.3
058% we; Esq - in: 2582 - 5.3: :98:- :~.3- 5:.- Su: r8-
383 380.0 3930 883 A983 8.80
:20 388. -298 -08.: -298 +3 .1; -98
$3.30 893 E95 3.80 883 3.80

59

.0050 —.0 608830 was.» Hap—3. 0..-30888.. 30:3

0.8% 0:00.8— _. W088

008800 mgr mun—.5. .. 0.8% 000000—
00888 mgr mop—5. .. 0.8% 98000— .
romA8m080

r0<08m0

8880;08— 8808

romAmmﬂ

H8820 88082-03: 8808
€553 08:82-08. 0808
romeo»: 88550
romcog 08055

m00500

<08 008808

8085. 0:888

r08 050080 008808

0802888
”-89.0500

r3931... ranges... 0290...... 0:3...
Ewan-$0 33.0: Awe-0.000 3.30
-mm..3_.;:.. -8001... 80.00:... -93....
30.03 B0,.— 3 AN _ .000 3.80
8.3 3.00 0000.. 0:0.—
3903 $0.03 @0900 8.30
-3933... L000?»- -0_.3 0..:
A00. .3 30.08 80MB 8.03
000. .0... r0300. MSU; 30:00
3.0.03 33.09 33.03 3.03
-.00_.~0:* 90.03.00:- -u.30.~_:. 0km
$3.000 3000.03 33.08 20...va
.0000
GWNS
30.0—
33.08
30.3
300.08
$0.3
Swab-3
-3080:-
$0.03
:5. _0
33.000
<8 <8 <8 <8
<8 <8 <8 <8
20 20 <8 20
03 03 0: 03
0.3 0.3 0.8 0.;

0.8.
3.0.3
-93:
A980
98
8.00
-98
8.03
.93
3.0.0
8.8

GPA:

0.00
2.08
-m.00_..
3.9.3
$.0—
3.0:

98
6.80
-98
A980
98
8.80
93.
8.90
:93..-
2.80
98

A503

0.8
C. _00
.0001...
8.0:
0.0—
3.000
<8
<8
<8
03
0.8

 

60

Haw—o Fa 8370338— H.438

3.5 838 88.-3.. comm-883. 83338 3.03 «own-.883 8.38m :5 338832. .8: 32888 8 8388.... 83892.38» 8.808. 858%
:33 .3830 836.383.. ado 8283383 818.8 38 3o Imam—3333......nm03333 88x .3 3n 08.333328 8.338 :33 3a 88. £88 0..
3n 8»: .88 3.. 8a .33 33332.3 .3 :8 «<0:-33302.8 8.3339 A...» «38388 :3 833383 3. 33.330. .83 3:38... army 0388.
.m m 833v. :3: 8.8m o: 30 «3.30 o.- o3a z. :8 3:: .m 33.88: 8 w 8:83.. 333 N03 33232.8. .52. 0.808. .m m 833v~ 33 8.8m 33 3a
<38." o.- o:» :- 3a 9.3 .m 33.883 8 m 83 120 888.. 33.. 88 0303.8. 3.8% 8308. .m w 933v. :33 8.8.: o: :5 <m.co o.- 33o :- 3o
ma: .m 33.883 8 a 303-83 98 0.888.. 33.. ~98 2.83.8. 0388. ”5.883 .m 88332. 888.5 8 30 .88: mm: H330
0033.383 GAB-O. 33.. Fan 38$. 93:: $58.0 8.38838 .5388. 858mm. 333 .6388. 883a 8.8.383: .3 18m. > :33 w.
«8.82.37: 0.583. 858% .m 3a 338 3228: 3a .8258... cm 30 0.508. :38 .8 88. 388. 0.508. 883.. 8:138: 8328 30
8.8.38: 302.83 30 8:35} 883.. :38 238302.38. bra-68. 33.. .6 33.88.. macaw .383» :38 8.8.8.5 ".5 8.8.22. «.38
.28. .883: .m » 3:33v~ 3m: 8.8m :3 3a 8.30 o.- o3o 2. 3» m8: 338 o.- ..5 8:: .m 33$ .33. 333 88 2.83.8. romAmmmaa. .m :8
88133 3.. Eo ma: 88. 838 33.252. .3 38-83% o». mean. .88»: «<33. 883%.. .m 30 3.38 cm 88. 88:88 8 88. 888.
.383oD-38. 8.88 .m :5 :38 cm 32 .383n 8 88. 388. romAmmo. .m 30 88333 3.. En 33332. o.- 383. £38 30 .3898: 3.. 3o 33.
romeo»: 3333.5. 8 9» 88133 c». 30 3333.5. 0»- 30 .83 8 38.? romeo»: 330333 3.. 30 .5333 on. :5 8:: m8».- .: Cm 8.83.
woo—=8 .m w 833v~ 33 8.8w 3: 3a 8.5 o.- o3o :- 3o .23 .m 893.8. :33 ~38 33032.8. 23333.. 8 3383382.. 3288 88m" $3383
282 0.888 a 90 an: .n<o.. ti. :- 33: .. 83.88 33.3.8.7. 38.283 3 =5 ...\e. uﬁ: 333 .o..\.. .30..

 

33o. >. 0.808. 558.9 0.883. 838mm 133a. m. 0.880. 558.9 0.808. .3830 038.28
Ina-33:3. r83 maﬁa $01.33»... Pam: mrwno 22:33:... r83 mrmna Ina-.38... F33 $88

 

 

 

A .. Q. G. 3. G. 8. G. a.
0380. red:- -3. .353: #3:
83.8. 3:. 08.3. 9......
0382.323: 3.32.3:- -._~ 23.8:- .8
.38. s. c 3... a8. 3. 3.3.
mini-98822330 3.398:- éS 3.3 pd
. :8 _ .3. 2 .3. 03.3.. 3......
98819.53. Essa-38333.3:- e... -83...- a...
:33... :98. 33.2. 2.3.
32.0308. $3.:- 3? .35....- -93:
35.8. 3...... 36.3. 3.3.
32. 2.88. - E33. 9......- -:.8 3.8. 3.:

3.3.3. $3.3. 33.3 603

61

.28. agave. _. 0:90.8— £53?

Haw—a 94

.28— nrmnca .. Orangu— ﬁimgo .. Women:

.203 252x:

633 2532 _. W828

ﬂew 2:560— . 9589 $8320

ﬂab 035g— . 035?: £535 .. ”$2.3

rowAmmmﬂmv

raga—mo

583838. 88on
romeo»: 895:3
romeoms 2:953
Mango:

<9: 955:8
5985. 9.398
Pom: 1:680 955:8

0802988
”.358:

L93

333
393
33.85
8.83.8...
3.38
-83
:35

53.5—30.5. Managua—.3

-9898...
829:
8.398.
328.83
83.8.

62.5

988:...
2823

$893..
2823

83.3...
2838

9.3 -u&.oa

8?: 8.8V
.93.
3.“: 82.3
r: 8.88.3...
33V
-93 .88
:.§ Sid

ugh:

Gwau 5

$383.: -93....» -oou.&o.:

A&w.&3
$93
Awaabov
<8

A39 3:3

90— 93

A”: C $3.ch
<8 <8
<8 <8
<8 <8
03 e3

98 PE

-28
9.5
2.8
6.23
.So
:38
-98
29.8
-83
:93
5.8
8.5

PS

865
-93....
3.3V
FE
CNSV
upw—
2.5V

5.3:...

8&3
a. E
3.23
<8
<8
<8
0 _ q
98

5A. —&

899:
23.3
A3 9?:
$58.3...
33.2:
553
:8. _8
83.32....
BMMNV
-&&u
€3.qu
<8
<8
<8
03
93

3:.
33.3
58. a
:3.on
-888.
85.:
28.3...
as. _ 3
~38
$58
-28. 6.
38::

9mm 5?:

86$ 39%:
.&.omu we? _ &
3&3 3%.qu

9G -u.$_.mo:.

25.3 $3.98
5. _N .moba
2.5V A5568

-93.... Ame—mm...
3&8 Sub:
_bN -&.ow
9.8 $238

<8 <8
<8 <8
<8 <8
0: 03
PH 93

-98
3:3
.3“
3.8V
-9;
93
98
0.5
NB
3.8:
-98
3.3V

9%
8.3V

Lo; 1....

3.8:
_.3
c 33
-5.
2.9

-932...

8&8
93
AN. _ 8
<8
<8
<8
2 q
98

 

62

Haw—o F: 883:. 2:— miaaﬁ 0:83.88

:8 8:8 883.... 882.883 8:328 $2: 888888 8:55“ So 3:92:2— 8»: 82888 8 883% «38:88 8388.
38:8: 3. 28: 33883: 8:88.88? 9:: 3. 8 58.88 83 :8 mm: 83.8: 838— :8 92883: 388. 4.8 5.8::—
mo<8:m:8 382.8 :8 882.88: m. =8 :28. 8.. En 8: _o 9:88—88 :: 55o— ? 9:: :6 8: m 9:88.88 8 :88. w. .28
88:88:: 81:28 m8 :8 18388372833888.“ 810:: 8:88. 8: =8 88. £58 cm 90 .8: :2: 3. :8 :38 88:823.
.38 89888:... :8 88:33 3. 82:30. .28» 3:28 arm: 0:88. 8 9 9:33. 9:. 8:8 8: En <28 cm 0:» a. :8 3:: :m
35:98: 8 8 0:88.. 8: 88 2:82:89 .501 0:862 mm : 95:3. 9:: 3:8 8: :8 <38 ca :8 2. :5 :3: mm wad—58: 8 w 8:
:8 888—. .5: 88 282.8. .283 98.8. mm m 9556. 92 8:8 8: =8 8.8 o». :8 z. :8 :3: :8 83:80: 8 w 25-8: :8
888.. “5: 88 2:838. 0:862 mam—8:8: mm .8258: 888:5 8 :8 ~88: mm: H88 0:83:88: §Q MS: 8
383. women: :m m 9:53. 2:: 8:8 8: :8 <38 on. :8 a. :8 m8: 888 on :8 8:: a :38 $3. 8: 88 028218. roam—888V
mm :8 _omﬁaz: on :8 :2: 88— 888 388:8: :: 28888 on mac”: ~88»: (<8? 8838 mm :8 3:: on. 8:: 58:38 8
88. 888. 38:82.08. :88 8 9o 3:: on :9 58:8 8 88— 888. PowAwmov 8 :5 8818:. cm :8 85:8 on :88 m::8 :8
588:0: om =8 32:. H8558 8882.08— 888 mm :8 8:: ow 858—0 888 8 88— 888. €855 828. mm :8 85.838
83.8: 258:: 888 m5: 958:: 53:38. r828: 33.509 :8 En 88155 2. :8 38:13. on, 5a 8:: m: :88. hence»:
3:253 mm 98 8819:. om :5 8:: :8 :: Cw 8:88. 98:8: :8 a 95:3. :5: 8:8 0: :5 8:8 om 8:0 z, 9o 8:: mm 888:.
2.: 88 058,288. 25:88 5 3:83:88 8:08 8:8" 35:8: 888 2888: m. :8 an: .88.. .... ... 8: . 8:38
$338.7. 85:58:: :. :8 _..\a. mob. Ea 5.x. _0<0_.

 

v98— ? Ho: Ho: m::_.n:o_88

 

 

 

283:8! 8:: $58 $838»:— Fom: 258
A: E 9 3 3 a: 3 3
92:83:: 8:83:58: -Na; _.. . -3. _ N. . -o. I. -o. _o

3.33 3.ro 888 32$
9:883: 8:83:50: ~93... 3.08.. o. 5.. c. _u

. W883: Anni: 3.1: :58 8.23 .

9838— 3.5283: mm: 0.2. New. o.oo Poo
3.8V $3.3 3.3V 8.93

955: 92883: m8 . 58:: .93 .93 92 o.oo
3b: 3.93 8.3V 8.on

romAmmmoav -Nmabo... Lab: -_ we. can Leudw... .880 -_.uo.. 9%

39¢: A385 8.9: Ba: Swag A308 8.2: 8.qu

63

798250 w 5.3 -_~_.uu -w.m~ LPNN: $3.3 Lara .q. 3 ._ _hot.
3803 33.08 3Q C 3.3 Soc.~3 3353 3&3 3&3
58302.2”: 888 b.mcu.w~:..u.moc.mc. .3. Lung. -Nmba -mbNNS _.:.:..m.:~. 8...... - 2.8 .Nahm
:aooud 203.93 AMP—3 A39: Cameo: Cgaawv 392$ :99:
romammov wwcboﬁ. .. rum 383...: foo
23.23 2.33 2.9—“: 2...”:
H2565 88082108— 8808 5:5 Lbo hob.“ L31.
33.qu 3.3V CANNE 3.3V
€9.55 SEE—2.0:: 8828 N. _.o_ -93 $93 -w._~
33.3 8.3V 3803 8.5.3
romeo»: 33509 .3.3 -93 -uobo .93
23.93 AZB Cure: 2;:
FOE—om: 9:555 -3931... -932... .aﬁbat... 560......
Gm. _3 8.3V 2868 8ch
moo—=8 L 3.2 :5 -Nwo. _N 9%
$3.ro BN9 @803 Show
<08 955:8 <8 <8 <8 <8 <8 <8 <8 <8
3985. 955:8 <8 <8 <8 <8 <8 <8 <8 <8
to»: 9558 955:8 Zo <8 2o <8 2o <8 2o <8
0824328 3: «3 98 at 08 «.2 Ba at
$3539 o. 3 98 o. _u 93 PG 98 PE 93

Haw—o Va :52...»— .:=_ min—.5». 0303:323355

 

64

5.0—0 :.0 5:030: 0:0 M03030: 63.03.0000 80:0:

:.::0 80:0 00:00.40 80330:: 808080 :.:.08 ~0mn000:0:0 ~0:0:::m :00 $508080 :00: 00.0088 8 5830: m0<030:00 0023—0.
080000 0v. 0800 0550808 00:00:88? 0:0 00. :0 800050 0:: :00 «0: 003.00: 000:8: 0:0 05.000008 :.:m0:0. :.00 8:030:
m0<0~:0:00 800058 00.0 88:80:00 0: :00 _0<0_ 0:. :00 8:0 :0 000800—080 :: 00:0: .0: 0:0 :00 8: a 000800—080 :: 00:0: w. 30
0003008810080 000 :00 :03:000:-::~00:=:0: 800x :.0: :00: 000000. 0:0 :00 88: 00000 00 :00 :00: 00:0 0:. :00 :000 0:050:09
H00 «38888 000 8:80:00 0v. 30508. :000: 000008 803. 000000: :0 0 008:3. :00: .008 0: :00 80:0 0:. 0:0 ::. :00 :.5: :0
030080 8 0 000000:. 0:0 008 0:003:00. :.:0: 000000: :0 0 0:803. :00: 8000 0: :00 8:00 0:. 0:0 ::. :00 :.5: :0 03:080 8 0 8::
30 0000000 0:0 003 0:002:00. :.:03 000000: :0 0 00880. :8: 808 0: :00 $05 0:. 0:0 ::. :00 :08 :0 03:0:00 8 0 :0:-8:0 :8
000000: 0:0 008 0:003:00. 000000: 03:0:0: :0 303300 00020::w 8 :00 :A0000 :000. .3000 00:08:88: GAE-O: 0:0 :.00
00-003. W082: :0 0 00880. :00: 808 0: :00 <0::0 0:. 0:0 ::. :00 08: 00:0 0:. 0:0 :00: :0 030:. :03. 0:0 003 0:003:00. 8080008:
:0 :00 :0m0_.::0:: 0:. :00 :.:8 88: 00008 8000500 :: 80:00:00 0:. m0§0 ”0:00: €08 :-0<08m0 :0 :00 00:0 0:. 88: :00:::::8 8
88: 00008. 50080:. 08: 080:0 :0 :00 00:0 0:. :0: ::0080 8 8:0: 800:0. P0m:0m0: :0 :00 _0m0:::0:.. 0:. :00 05:00: 0:. 000.0 0800 :00
0:00:00: 0:. :00 :8. :.:—5:08 0000:02- 08: 0000.0 :0 :00 8:0 0:. 8:m:0:0 00008 8 88: 00008. €255 000:8: :0 90 05.8080
002.00: 0840:: 800:0 0:0 050:: ::00::::8. :-0m:00: 80813:: :0 80 :0m0=.:08 0:. :00 8080:? 0:. :00 :00: :: 0000. 08:80:
080:::: :0 :00 88:98 0:. :00 :00: 0:00 :: Cm 00:000. m00=R0 :0 0 0585. :00: 8:80 0: :00 <0::0 0:. 0:0 ::. :00 :00: :0 0000000.
0:0 003 0:003:00. 2080000 :: 0083088 00:08 8000: 0050000 02000 080830 0: 80 :00: _0<0_. 3...... .3... 0:0 .. 80:00.8
08000005. 05:38:: 0: :00 :00. 0.8. 0:0 3.x. _0<0_.

 

00:0: w” HoFmWE m00~000:00~0

 

 

 

0:03:80: P000 m00~0 $035000: P000 M0000
:0: :8: c :r :0: c0: :3: :0: :3
92:80:00 0080:8000: .200- -:N.00... -92: -93
3.08 :93: :900: :93:
0058008 8000:0000: ~93: :93 92...... 9 :0.
.. 90:03: :93: 3.3: :903 :908
000:8: 03:80:00 m0: :93...» :93:- 900 900
3.00: 3:». C :99: :99:
00:30: 08800:: m0: .. 90:03: .5901. 90.0. -93 .98
0.00: 3.00: :93: :99:
0580008: .3930: .0000 -90: :.: :. .3932; 0.0: -93 :.:0...

3000: 3903 8.03 :90: 3903 A098: 303 :93:

65

<00—0 :.a 5.2.00: 050 M0830: 03.03.0000 :00:..:

r0<0q0m0 :. :::N.::~_.. 30.00
300.00: 3093:
800802. 0.0: 0000.0 .9009»... a a... 30.8... ._ ..
2.0300: 2.08.2:
000:000: :0 : .00
30.3:
.:,0:m:0:0 0.000% 0.0: 0000.0 3.0.0::
300.00:
5.01:5 00385. 0.0: 0000.0 03.5
3093:
rom:_00: 50.53.: -00.w~
23.00:
:.0m200: 0:88.: 000.21....
993:
m00E00 :9 3
$093:
<000 0:88:00 <00 <00
50:05. 0:88:00 <00 <00
:.00: 050000 0:83:00 20 <00
000020.83 00:: «.3

9.090030 9 :.: 9.0.:

.000
3.00:
.300
20.00:
_. z
8.3:
-03
0.00:
0.8
3.2.:

-3000:
:93:
:92:
::~..:.::

-00.:
2.8:

-0091...

:93:
93
2.0::
<00
<00
<00
mum
9N0

03b: :.3
300.3: 30900:

-u.mu::.c.:......u.::0.:.u0:a

2390:: 23:05
3900...

20900:
-890:
3:93:
”.0098
:0300:
-0900
::~: .NA:
-393...
@900:
090:
:3 : :3:
<00 <00
<00 <00
20 <00
00:: 3::
9 :.: 98

-000
3.0::
-000
20.00:
_. _N
2. _0:
-000.
:.::
.000
3.00:

<00
<00

00::
9 : w

:93...
:93:
$900
::w.~::

-9;

::.::m:
-0902...
:900:
:.m:
:Nbu:
<00
<00
<00
mum
9N:

 

66

ﬁne—o re 88......— 28 H.326... 0363.58 :— 0.80.8...

5.? 820 88.8 08392: 83828 2.3:. 88888.5 8838 :.o $38.88.. 8m: 222:8 8 0:88. 93:28.. 8:858 82 58:.»—
28 983». 8832.8. .28 88:82.. $58.8 ”.8 8a $238.5.-Eanram: :82. 82 82. 82.8. E... Ba 8:.— mrmno on, :8 82. 83
3. 90 .28 23828.. .58 «8888.... 26 82:88.. 3. 035.5. .25. 3:88 AOer 0.528— .m a 95:3. =8. 8.8.. o: :8 <88 cm can 1.
90 an: .m ”5:888 8 m 2.88.. 2... 88 080358. .22— oswnci mm m 95:5. :8. 88m 9. =8 <38 on, 08 a. 80 :3. mm 8.8.2. 8 m 8..
38 2.8.8.. 28 88 0822.8. .503 2.88. mm m 95:3. :5. Exam 0.. 9o <18 a». can a. 50 33. mm 8.88.. 8 m 8.78.. :8 2.8.8..
28 83 28238. 9.88— .5083“... .m 82:52. 882:8 8 =8 .88.. mm... H88 0888.88: 00.40. 2:. F8 8883. 00832.8
888836 28w 95.288 8:03:28: 8 ESQ >. 28 .8 mm“. 8328.. 838. 28 9288...“. 1%.... 8 3.8— w. ”263. mm a 9.8.3. 8m.
Exam 2. .8 5.8 cm 28 1.98 MB... 88 cm .8 82. mm .52 88. 88 83 0283.8. r8882... mm 88 8818... on :8 an... 88. 888
.8852. 8 288.5% 0». won... .888. €03. 8888 mm .8 8.8 om 88. 52:28 8 88. 883. 885.82.08— .58... .m =8 8.8 3. :2
88.8 8 88— 888. 8888. mm :.o 8818:. 0». :.o 3.582 a». 88:». 8.8 80 8838: o». :8 ma... gamma—o 8882.08. 888 mm .8
8.8 cm 8888 8.0.2... 8 88. 888. €9.85 26:..— .m .8 8.3.22.8 838.. 95.2.. 886 28 952.. 52:28. romeo»: 82:10.. mm
:8 8818.... On :.a .8813. o». .8 82. m: 885. romeo»: 2:95.. mm .8 8818:. am 8o 82. £8 E Cm 8:3. mania mm m 88.3. 8m.
88m 2. 50 <28 o». 08 2. 8.... 82. mm 82.8... 28 88 8803.18. 2.5.83 8 882.988 888 895. £288... 88.... 2.5.22. 2 80
8.... 88.. 2:... .2... 28 .. 88.88 mgaaoaq «.858... m. :8 1x: m2... 2:. 8.x. .88..

 

3.8. >. 00832.8” Eon—n 0.5.0338 08:83:28..

 

 

 

$23.88: 88 2.8 Ewing»... F28 £58
2. AN. 0. 3. G. 3. Q. a.
00832.8 -88....» -388: -o. 3... -o. 8... .888... L981... .8. _ 1. 8.8
8. _m. 8. 3. 8.8. 8.8. 8.8. 8.3. 8.8. 8.8.
09832.8 .. .883. 3.3:... 8.8.... 8.8: 8.5 8.8:... 3.8.... 8.; 8.8
8.8. 8.8. 8.8. 8.8. 8.8. 8.8. 8.8. 8.8.
9.88. 83.8 83.5... -93 .93...
88.8. 88.8. 8.3. 8.8.
0:88. .. .883. 88.8.3... «88m: 98 P3
88.3. 83.8. 8.3. 8......
d2.— ormoca 5.3.8 88.8.. -9: -98....
88.8.. 88.8. 8.8. 8.3.

67

do: 253°— . ”2,2,3
10% gnocc—

donw nuance. _._ wnmoad
SWAP-82$

553%

:8anth 8828
romﬁmav

ANSWER 8moBD-o.m_ 888

$5135 8385.2»— 888
romeo»: =52:va
romeo»: 3555

wanﬁna

<8: 9538

~385- 95:38

Pom: 2588 95.38

0802308
w-MEmBQ

Hue—o Fe 5.8:...— EE Mano—.5.— Dc<2dn=no ma ﬁancee—23:3

-8931... L93
393v 39.9.3
35.8. 3.3

3.9.8 3.2.2:
-9398:- -9295:-
wawuov 23¢.ro

53.3.:

2368
3:3
3388
m8. 3
33. _8
89.:
23.5
53:81:.
A238
-325
3.358
<8 <8
<8 <8
20 <8
So ma—
o. 3 93

L8...
8.5
#3
3.83
-5:
so»:
58
:83

:93

308
-.3
3.3V

<8
<8
20
So
a. _w

93
8.5
-53:
3.3
-32
:83

.93
3.7:
$01:
8.3V
_.-
3. 3V

<8
<8
m2
ohm

mar-Kt. .293.
$3.08 $3.23
$50.3 -393
6258 $3.53
r2931... rowwuu:
3o _ .e3 3833
-mwm. 3.1;. burg
GN-V: Amramv
30.3 3.3
33..on 3.9.9 _ d
8,3..82... -u.wmu.cu:i
23925 22338
maﬁa:
2238
38mm
03::
84.3
33.03
#93
ASN.3:
$5.83..
Gab—V
Lwﬂum
@8th
<8 <8
<8 <8
2o <8
03 a...
:.: Pg

8. _N
3.5
-_.:
0.8V
88
3.3
-25
8.8:
-82
3.3

-38

so.“ 5
98
:83

hbq

3.8V
Lam
3.03

<8
<8

Sc
9;

T:
3.8:
-88
see
Bu
3.3V
9.3
8.5
._ _3:
9.:
3.3
:33

-o.mw
2.3V
-981...
8.3V
:N
S. _ 3
<8
<8
<8
wt
93

 

68

Haw—o Pa 383.... 23 mans—dc. 03.2.5239 :— nrno—S—u 32.3

:5 82¢ 8.83.. 80392: $3589.. 368 Rmnnmmmosm «23:5 :5 358323 8»: $2088 8 nance. ”—35.8: 8385.5 8a 883»—
25 38:5— mo<n3manﬁ .25 83033» <24»an 98 5a Enamnamrmemogm—s :59 m: 8a: 328. Ba :5 88— mrmna cm :5 8m: :29
3. 9a _25 magmnzwv. .25 nomammmoa 3o 8:32am 3. 0385 38.: 35:8 army 9:63. a a 95:5. 5»: grow o: :5 £95 on 25 z.
2.5 :3. mm mam—589 8 m orwncor 88 ~28 8302mm? in: ovum—x.— mm m 9556. 92 grow o: :5 $55 cm can a. :5 3:: W 93:28 8 m 8b
95 2588.. 25 N08 oﬁagmmn. ﬂQN egoce— mm 9 95:3. 58 8.8m o: :8 <35 2. can a. So and mm ”3:23 8 m =o:.8u 35 2525—.
mag No8 892.53. Q538— mazwaoa mm £25an 88235 8 Ea ~83» mm: ,5on OoBBMmmmoa A39 88 P8 38$. 0853958
33825.. 200* 9230876 8:83:30: 5 35a— ? Ba 90 mm“. .3300: 8:22 25 osdoaau 1mg. 5 351 w. ”2.2.3 mm m 95:5. :5"
Swan 0: :5 5:5 2. can x. :5 mg: 38 0». 9o 8»: mm M52 33. 85 33 825213 romAmmmnBV mm :5 _omwazs. on, 50 an: 88— 888
52553 5 28:32.» on mos": K03»: $2.. 7053mm mm :5 3.8 on 88— .82.53 8 88— 8.88. 38502.88. P368 mm :5 8.8 on :2
58:5 8 88. 8.88. romﬁmnv mm :5 _omwazﬁ. cm :5 3:362 ow 83a £38. :5 5838: cm 90 :3... H2583 8882.9”: 3.88 mm :5
8:0 3. 8:920 883 8 88— 888. £355 838— mm :8 93.9.38 32:53 952: 2.308 mag 9:42: 59:28. romeo»: :88an mm
:5 _omunzﬁ. 2. 9m 3255. an :5 8m: 5 88¢. romeo»: 88.53 mm So _ommazds 2.. 9n 8»: .98 5 Cm 8:»? wooEaa mm m 9.53% 9m»
Exam 0: :8 8:8 ow can 3. :5 8»: mm 39:8. 28 N28 8503mm? 25:33 5 62.03583 8288 89E .6533 283 0.58qu 8 :5
an: 65.. 1:... 2.. Ba _. 5&88 £3335. mmmamoma a :5 1x: m3. BE 3.x. _o<n_.

 

35o— w. 0953889 02:8— 96

 

 

 

=n~m=a§_ roam mrﬁo moan—am:— roma maﬁa
A8 :8 c c :8 :3 :3 :3 :8
032258 Hue SN“? o.oc obo 4.3 $3... -93 98
gay $.qu 8.3V 3?: $.va ENS 8.9: 8.8V
039.258 ._ women: .9; .93 Pa 93 .9; .33 93 93
Gad 39: 25.: 8.3. Gar: 3.93 8.9.: :53
arm—aco— -umu.mo -393... {Poo -93....
Augﬁv amour: $.an 3.8:
0:862 ._ ”2.2.5 «8.3:... «$.33. 93. Now
0588 0288 $.qu 3N3
do: nuance. $3.3 -3??? -98.. .932...
Gav—av Gucmmv 0.2: 3.83
.22.— oruova g women: 48.3.... 39:... 93 73
Ammo... : Goat: AP— : 6.5.3

69

d8~ arm—aco— -m~.um 82.3 -_ be -93.
GAP-:3 3%.qu $.me 3&8
HEB argue. _. ”03:: 788.23....- rcucbc: ubc u. 3
$3.on 3383 3&3 3.3V
romAmmman $3.832. $3.? -73 o.oa Lorwat... -_.mu 5.: r3
3 _ .5 Eu. _d 8.98 8.5 QN§ 895V 8&8 8.93
r9525» 3 _. _N uuuo 5.3 -93 A 5.3 Loam -o.- .93...
A: _.qu 38b: 8&3 3.3V 3 _ouwv 33%: 368 3.ro
FSBoh-BE 888 -u-Aowuuig 5.3—.33... Lt: -u.um $155.3:- -u.uuo.3:. L98 bf:
Swab-3 :43.qu 3930 Con: CNN-Sm: 23.9.93 ﬁsh-av 23.8
romAmmov mocha-2.... #3. Egg..- fww
ASP—8 :.qu :3.qu 2&3
H8535 8882-2»— 8808 89% LPN» 3.3 b. _o.
33.an 3b.: 3.2wa 3.2:
ﬁeanm $385.99. 8828 39% F?— aomAN Lba
33.8: 3.3V me.ud 3.3V
Fence»: 535.5; -uw.wo -93 LP; bum
ASP—8 :.Bv :3.ch 2.5V
romeo»: 2:055 -owa.~_t.i 5‘33... 53.3:- 53......
22.55 8ch 2233 8.3V
monﬁoa .3 _ .: o.om buobu Poo
€3.me AN. _3 ﬁg _ 33 AN. .8
<8: 955:8 <8 <8 <8 <8 <8 <8 <8 <8
5985 95:38 <8 <8 <8 <8 <8 <8 <8 <8
be»: 3680 9.338 2o <8 2o <8 20 <8 20 <8
0802288 0 _ c at o 5 at o _ e mi c _ c at
w-mncmana c. 3 98 o. 3 Pub c. 3 9.8 c. _ a chm

an...» —.w 3.2....— ana 5:2...»— nocoansno m: 0.50ch Ana—.3

 

70

Hue—8 F3 3.2.88— 2:— mks-88.03.288.881: 0888—88—8

.288 888 8885 88838888» 88388888 3888 889888888 .8835 98 $5888.88 88: 88888.8 8 8888.8“: m8<838888 88888—8. 2.8888 3.
888w 82888888 8888885888. 888 3. 8 8888888 83 88 m2. 882888 88888— 888 8888888 1mg... 8.88 888888888 88288—8... 28 58
I838amEimaor—888 88888 88 888 88888 8 8888— .? 888 8888— n. 98 88— 888.8 8». :88 88.8 :88 3. 58 _888 828858sz 8 8888— w. 888
8888— U. 888 ~8m~88888m 2.8 888888888 3. 82:85 88.8 38888 Aormv. 0.8888— mm 8 88883. 888 £888 88 98 $888 8m 888 z. :88 :88 mm
83—8888 8 8 888882. 888 88 8882388. .28: 888882 8.. 8 98885. .888 8.88 88 88 $888 88 888 1. 98 mm: 8.. 83—8888 8 8 88 38
888882. 888 N88 88:82:88. .283 888888— 88 8 95:3. 988 £888 88 98 58:888. 888 z. 888 :88 88 83:23 8 8 888.88 :8 888882. 888
N83 88883888. 088882 83—888: mm E83588 88888885 8 :8 W888 mm: 8.388 0885:8888 GOV—.8 888 9.88 388$. ”88:8 88 8
8888888 98. 8:88 88 98 «.888 8.. 888 a. 58 88: 88.8 8m 58 _888 mm 8288 $3. 888 88 8982388. PomAwmmoav mm 98 _8m81888 8m 58 3:8
88— 88888 8.888588 88 98888888 8m moan. £88»: £88. r8<83m8 mm 58 3:8 88 88. 58:38 8 88. 88888. 5888.82.88— 88888 8.. 98
8.838 8.. 888 8888.88 8 88— 88888. romAmma mm 58 8%:588 8m :88 85882 on 83.8 88888 88 88888888 8». 98 3:8. H§m§8 888885.88—
8888 mm :88 3:8 8m 88588—8 88883 8 88. 88888. €2.83 8888». mm 98 8388888 883888 88:88" 8.8888 888 88:88. 58:38. homes:
8.88509 88 98 _8m8193 8m 98 888810. 8m 98 _888 88 88mm. romeo»: 288883 8.. 888 _8m588 8». 88 _888 8888 8 Cm 88:8? m888~88 mm 8
98885. 98» 3:88 88 98 $.88 8m 888 z. 88 888 mm 8885.88. 888 83 8982888. 258888 88 888898888 888888 8888» 888883 8283
88888.88 88 98 :38 _8<8_. 1.... t... 888 _. 888888 88888885. 85838888 88 98 1x: uﬁe. 888 3.x. _8<8_.

 

 

 

+8338:— roma $888

3 3 83 SF 3 A8 3 3

0588878 88888888888 :.um 95 9mg -Zbu 9N8 95 9 _u 98
298V 29—3 A388 :89: A919 8.3V 8.3V 8.3V

0288888 888888888888 .. 88838 -93 Lou Lba 3.? -98..» -93 -93 .95
:8. _ 5 . 28.83 3.33 2983 A9 .3 A9~3 8N8 8.3V

9.8888— .593 -3. _w -_ F3 -393 .93 _.No 93 -93
$3.88 3.8me 0.388 3258 Ade 3.83 683 3.3V

988888— .. ”8888 $93 88.2 $2.3 «3.2.. .9: -9: .93 -rd
33.8.: $3.83 $9.88 33.. C 3. _ 3 3.83 1:8 8.8 5

0588.888 88888883888 .. 988888— -umbu. .w _. 3... .N8. 5.. -93 .93.... .93.... -9”: -9NN
A323 2988 :8.ch :9qu 29—8 8.23 8h: 8.5V
9288388 88888828888 .. 0:88.8— _. W883 3.3.. 3.2... $.88... 8.3 93...: 9mm... 93...... 93...
$983 3. M8 ANNN : 3988 8k 3 8&8 9.8.3 8MB

71

Hue—o r3 58:3— 2:— 92032 032.558 E Gama—328:3
romﬂwmmnav L36... -5933... #95 -obm ._ .Nm 9.3
SNAXV 3063 25.3 8.qu 8.2: 8.48
r9838 3N3 mambo 3.3 1:3 -93 L93...
Qﬁbwv SwNLE 33.qu 3.5 3.23 3.98
_snoaahdg 8828 ..Mo~q.~w1..u.u~m.au:39393:... Lube . 36w. -~_.cw
Saawbs Emmabmv 23¢.qu EPNS 995 Sub:
romﬁmmov Nmmbu. .. _.~o
Scab—v 20$
Huang—a 8828268. 8828 .39“: -93
3238 3.8:
«Sinai 3985.08— 8an aoohm -N. G
3368 3.8;
homes: 398..va -498 .93
2363 :.:.v
romeo»: 8:956 $5.31.... -938...
$39: 8.qu
mnoﬁoa Luwhc Pea
Emmbuv 3&3
<oB. 9.83% <8 <8 <8 <8 <8 <8 <8 <8
5983 955i <8 <8 <8 <8 <8 <8 <8 <8
Pom: .5380 95:3 .73 2o 20 <8 20 20 20 <8
0682308 SN 3o 08 m3 2» So 03 m3
”-558; o. _m o. 3 c. 3 aka a. G c. E o. 3 ohm

 

72

8.8.0 F... 8838. 88 Mano-.8. 038888 8 0.80.8.8 .83..

.288 88.8 8.88 88338. 88888 «88 8m8888 8.88m 88 888.888 .08 8.888 8 .8888. 88888 888.8. 2.988 8%
8.8.» 0888...... 8:8:Q88? 88 8% 8 808:8 0.. :8 m8 .8388 888. 88 08888 188. .28 8.8888. 818.8 88 88
285888.-:.~8:88 .889. 8a .88 8.888 8 .88. >. 88 .888. O. 88 88. 888 o.- ..8 .08 8.8 8% =8 .888 888818. 8 .88. .w. 88
880. U. .38 888888 88 888888 8% 28.88% .88. 8:88 .05.. 088.8. .8 m 888% :8. 8.88 o: 88 8.8 o.- 28 z. 88 m8 .8
83.888 8 m 88.8.. 88 88 883.8. .28. 88.8. .m a 888% 8.. 8.88 o: :8 8.8 0.. 28 ... .8 8.8 .8 83.888 8 m .8 38
88.8.. 88 88 2.83.8. ....8~ 88.8. .m m 888% 88 .88 o: 88 8.8 0.. 08 R 88 m8 .8 83.888 8 m 88.8 :8 888.. 88
88 3.83.8. 0.880. 83.88: .8 888.88 888.8 8 88 .88» mm: 8888 9.88.88: cad-O. 88 C8 388.. .888 .8 m
888% 8.. 8.88 o: 88 8.8 0.. 88 ... 88 85 888 cm 88 .08 .m 8.8.. $3. 88 88 8.83.8. roam—888. .8 88 .888: o.- ..8 8.8
88. 888 88888 8 .8888... 0.. m8... .888 $8. .8888 .m 88 8.8 0.. 88. ..8.....8 8 88. 8888. 88882.08. 888 .m .8
8.8 0.. 8. 8888 8 88. 8888. F888. .8 88 88:88 0.. 88 8888 0.. 8% 888 88 888.8 0.. 88 m8. H8888 88.508.
888 .m 88 88 0.. 88.8.8 888 8 88. 888. £8.88 08.8. .m .8 85,8888 .8288 8.8:. 888 88 0:88 ..8.....8. romeo»:
88:3. .8 :8 .8888 0.. 88 8.88% 8.. 88 88 8 88%? romeo»: 888. .8 .8 88188 c.- 88 88 8.8 8 Cw 8.88. 8888 .m m
888% 88 8.88 o: 88 8.8 o.. 28 .m .8 .08 .m 8888. 88 88 8.83.8. 2:888 .: 888888 8808 888. 8.8888 888
c.8888 w. :8 88 .88.. 2:... 2.. 88 _. 88.88 8.888.? 8.8508. 8 .8 .98.. 8.x: 88 8.x. .88..

 

 

 

.u8a. n“ Ema-888. .888 M888

6. .8. A... .8. :8. .3. .3. :8.

0.88.8. 8.38.8... 8.2.3.. 88.89... 8.8.8.3.... -8... -8...» -88.. 9.8
3.8.8.88. 3.3.5. 3.2.3. 35.88. 3.88. 3.3. Cu... 3.3.

088.8. .. .888 .8381»... .h88.qm..:......mu..3!.i .8893: .93.... Jam: 3...... 0.....
0.5.3. 3.8.3.. 388. . m. 38.3. .88 .. 3.3. 8.3.. 3.3.

08.8. 0888...“. m8 -9”... 8.8 by”... -.uo -93 -93 -888 -988
8.3. 3.88. 3.8.... 3.3. 8.80. A... .8. 8. .8. 8.8.

0088. 98888 m8 .. .888 «.88 0..: :80 $3 8.8... 8.8... 8b.... 8.5..
G. .N. G. .N. 3.0.. 3.8. 8.. .. .o. . .. 8. . .. 8. .o.

98.8. 08888 m8 _. 0.880. .9882. .888 3.3 .98 8.88 8.9. 8.3 8.88
3.8. 3.8... $.88 3.0. 8. .o. 8.. .. 8.. .. 8.80.

0088. 0888... m8 _. 0:88. .. .888 $8.31. 3.2.8.... -388... 8...... -98... -o.- -93... .9...
2.8... :.bw. 2.8.. A...~.. 8.... 8.3. 8.8. 8.5.

73

H8... ..:. .8838. 88 H808». 03.08800 8 08088.8 88:...

P888808. 8.3.8: $3.82.... $8.88 -. . . m .. be. 8.8m
38.8. CPS. 38...... 8.3. 8...... 8.88.
.-0<08m0 .3qu $8.8 -388 -888 $.88 8....
3888.. 388.3. 388.88. 3.8. 3.88. 3.3.
888888. 888 .88. ...m...u..u.w-Bbm:__.u.mw~.m8._3.. 8.88 88.8 -38.
28.9.8. 2.3.88. 2.38.8. :98. 38.8. 28.8.
.8880. H.981... NS...
22.8. 28.8..
H8880 88.83.88. 888 .388 8b..
33.3. 3.8...
€2.85 08.82.88. 888 3N2. L...
33.8. 8.8.
.8888 888.8%. -38... .88.
2.3.88. 2....
.8888 88:8. 88. .81.... 8.88:...
Com. .8. 8.8.
M0888 .38.... .88
33.8. 3.5.
<08 888% <8 <8 <8 <8 <8 <8 <8 <8
883% 888% <8 <8 <8 <8 <8 <8 <8 <8
.88 8:88 888% 28 28 28 <8 28 28 28 <8
0882888 SN 8.8 e... .3. EN 8... 8.8 «A.
”88888 8. .u c. .u o. .8 93 c. .w o. .8 o. .8 8.8

 

74

.22 .8: :3 2n .2. 2.. a .5828: 2.888:
082:... .. :5 .2. .3; .80....0 8.8.0 82.. 8.. 92.8880 228 88:58 .822 088: 882822. 8 20252 .023828 28
:5 82:80 2 5o. 02. .. 28 .8 0:73 2: .8 80.8 82. 288:: 8 2 ..:2:00m.. 58.8: m... 8 02m 5:. 02. 8 82.58. 2: 2 .885
52.8.. .28: 2. 5o. 02. 8 8.888 2: 8 82.58. 02. 2 $2888 52.8.. 8:...8: 8280 :5 2088 802:0 803.2. 0038...:
02. 2 .8880 8.0.83 .2088 88. o. 2088 0.285. 8 0:2 2: 2 2088 88.8808 0.285... 8.: 02. 8 8.802.. 02. 025 2.8:
8 828:: 2: 8 82.58. 02. 2 88.8.. .2008 .88. o. 088:. .0: 8 0:2 02. 2 2028 88.8882. .228 .88. o. 3:...8: 88. 8
8:2 02. 2 0888.. .83 523. 280m 8 358:2. :. :08808 2028 .88 88.. 02. 8 82288. 02. 2 8.2.8.8.. .388 00.. :5
6.25.. 8.8.8880 0:2 .. 5... 823. 2: 9 28808 :0:.80:. 2 8.82:8 82820 .023828 20.2 :5 82820 a o. :08..:.8 2
8.: 2: 2 28 8 0:8> 2: 8 808. 85 288:: a 2 82820 .620. 88:8 .80. 58:8 3 :088:8 08 28.0882 02 .. 28.88:...
>50... 8:28.: :5 8:28.... .w8>m5mm. 202820 05 >02. :5 8080880 2.8.28: 8.02 8 :0:.802 05 20.23 m8: m. 8..
80.8208 mm. 08 202... 80.82.28 50. 8 80.. 2: 8 «23. 8 22808.4. 88.82 2: 8 82808 8 2 .o\o. 52. 8w8. 2 8.220230
3.80 82.22 :023 .28 880 2: 8... 8:: 882...: 8032.2. 2: 8 20:82 082... 2: 8 28 .823 28 8 0:.8> 2: 8 88.
82. 0.88:: a 2 ..8.8288 80.28.... .82.. 2: 8 0282.8. 8228.. 8 5o. 8 280:8 2: 8.8.2 2888.82 8 388:3 2:038
D .05A. .28 :5 28 2003.02 8 88> 2: .3 :080382 08 23.855 :80. 2: :2 :.02 5o. 2: .8 082m .82 2: :5 .882m
50. 8.. 3:2. 58202....28250... 2: .3.8.8> 80:22.0: 2: 22.3 .088 28. 88:02.. .3 888:3 08 20.23 28.8082
.8 2.82 8.88:3 2822 U .05.. .320. 38:3 .80. 58:8 .3 888:8 08 28.8082 2.... 20:5. 8 828:: 02. 2 Z 2023
.Z.-cooc.\x0::. 5882.34.88.20... .8888. 2: 2 22:80.8 :208280 2 80:22.0: 32. 02 ... 80:8 58202522888...
2: 8 80.2.. .0.8.8> 80:22.0: 30: a 95: 28.8082 8 888:8 882m m. .05: .Amqov 38:8 50. 58:8 .3 888:8
05 28.8082 2.... .023828 28 :5 Koo. 8.8 382.28 50. 02. 523 28 .8 88> 02. .8 80.8 8.: 288:: 8 2 ..88..0~...
...8.8..0~... :5 :.02829. 8 8:888. 2: 3o: 2 .205: 828 2: :o :5 .< .05: 8 28 18:85.. ... _. 82820.. 2.8882 ..288::
08. 8.82820: .038 2: 8 0.8.83 8:888. 02... 8285.8 :80. 2: b. :.02 5o. 2: .8 082m 88. 2: :5 .852... 5c. 8. 8:2.
582025258503 2: 08 8.22:3 80:22.0: 02 .. .023828 28 :5 63. 8.8 2:000 8.82.28 50. 823 28 8 02.3 2:
8 30.8 20.23 :85: a 2 ..8:.m5:-.3.... .023828 28 :5 .3885 .waa. 8 50.. 2: 8.8: 2:08 8.82.28 5:. 823 28
8 88> 2: 8 80.8 20.23 288:: 8 2 ..8:.m5...... ...8:.m5:-.3a.= :5 =8282U= 8 8:888. :5 3.8.2055... :5 .82820:
8 8:888. 888. 8:888. 93 :8. :0:.>.: 2 .2283... :5 .8282“... 8 88. 8:888. 2: .... 0.8.: 8 m. :5 .< .05:
8 28.8082 2: 02.8: 20.23 2. 28.8082 8 888:8 2:088 < .05: 28.828. .822 80. 8 388:8 28:2 0.28 22...

23... 805823. .... 0.2:...

75

 

 

 

 

due—o F: wow—.8388 Hanks—.5
was». > 55o— w
Iowa—am:— rama £53 202 oosomaqgo: was“
A: A8 A”: T: A3 A9
0532 LEV? -moobo. Lho Law: -92.. .98...
38.93 33.53 3.03 3.3V 8.8V 8.8V
950g. .. .330 9.35% .rngo. $89.3 9.2 -98 0.3.... o.oo:
A3303 A3393 8.8V 3.03 Aobwv 8.8V
9530. . 58755830: r3921: 398...... Sea: Pma
G :33 03.5 $0.: AWN—V
rcmAmmmnav .898... -39». -73.... 93 o.oo Poo
3.523 34. T: 868 8.3V Pcc o.oo
ro<8mmo r3981. Amu. _m Lbn L38. o. 5...... o.oo:
3.353 35.9: Aubuv $.qu 8.3V 8b.:
5830qu 8828 .u. Su.~c:_..u.amo. 5...: who 6. 3 aka: 5.8:...
:93. IV 6.363 :92: 2.939 8b.: 89$
romAmmov 3:3... F3 98
39?: 2.8V 8b:
Assad—o 88282.08— »mmﬂm Gmhq #3 Peat...
3mm. _3 3.23 8.89
€2.55 8385.08. 8828 34.3 L. 3 93..
GE .5 3.8V 8.93
romABuzin -wmb— -c. _ _ cbo
23.03 AZNV 8b:
romeo»: 2:253 53.21:. -932... 93....
can 5 8.”: 8b:
moon—6a _obo —. _N -93
Su _ .on ANSV 8.8V
5882.: Paw _ .q _ 2. 49393.: 3.33.. .. $98!... .98 -o. E
23¢.qu Sumcbav 25.98 :.ﬂmd 8.3V 8.5V
<2: 95.8w <8 <8 <8 <8 <8 <8
39.3 95:3 <8 <8 <8 <8 <8 <8
Pom: 9.308 9.55% . Zo <8 20 <8 20 <8
0802388 £9 03 £3 03 emu o3
”89:83 o. _ u o.No c. 5. 93 o._ _ o..—

 

76

Hue—o F: ”38388 H88 Boa;

 

 

 

 

551 O v25. U
Ewan—Ema r98 2.58 283:3:— ana $58
A: mp A8 :8 b: :8 :8 :3
03532 -93... -93.... -93... .93.... -uowbw... -oﬂbm: .93 -99»:
8.7: 8. JV 3.3V 8.3V SENS 3.383 3.3V 33.3
9532 .. i=8 95:3. conﬁ: 93.1.... 93.... 9?. r332... «8.3:... Swot. Pam
2:3 8. .3 8.23 8.23 $3.23 $3.83 308 3.3
03860. .. Panam— ooasaonoa No.33 Nmobo 5.8 :a
Amuom C 3.363 Bbwv 2.3V
0:860. _.. $0.33. 025832. .. women: .393 22.3 93 HM»
3303 35mm: 3.qu 8.5
rowawmmﬁmv -931... -98 -93.... 98 $8.31.... -333 .. _ .wm: ou—
Sbuv SbB 8.3V 8.8V ENE: 3903 8.3V 36.:
:8an :.:... .. o.No -93 -93.... _. _ m _ but. wmobw -93 -o. _ o.
8.~& 3M3 8.8V 8&3 32:3 3.9.?qu bev 3.»:
58802.23 888 ...um..:.:.. LQwSI 93 5.3 -u. $891.... 8.3931... Sb“ 9?
Ba: Ba: 3.3V 8.3V Cesar—3 $25: . :92: ENE
romAmmov P3... PE 73.3... :3
8b.: 8b$ QMNB 2.3V
Anzac—o 8882.99— 8808 o. S -92.. Nowbm Lb.—
8.~8 89:3 3303 3.3
£0135 ONES—3.08— 8808 o. _m .93 3m. 5 L. _N
893 8N8 $3.on 8.on
romABmeava -9?“ .93 -923 9.8
89$ :59 23.8: :._8
romeo»: 8:953 bunt... 5.3.3.... 58.31.... -93.:
. 8?: 8.3V @9qu 8.3V
moocnoa o.oo 9cm -u~.3 Pan
8. _ 3 3. _8 3393 AN. _ 5
58.609 -93 93:... -9: 5.3:... P8931... 3.48.32... 3.32.... 593...:
863 8.3V :39 8me A 33.93 20353 A 3&3 quuv
<92. 95:3. <8 <8 <8 <8 <8 <8 <8 <8
5985. 95:3. <8 <8 <8 <8 <8 <8 <8 <8
row: 9558 95:3. 20 <8 2o <8 20 <8 2o <8
0.862288 o3 o _ q emu c _ q 23 o _ q 03 c _ q
Wénlcmaa - - - .. o. _u 0.3 o. _ u chm

 

77

:.:—.8 :5 F00: 0......881888 an: Dana—00:...

.28 8:08 388:8 0083085 8:358 $08 8m8888 5.38:. :.:n 0:860— w3_8:0: 59:. :00: 050808158. .28 88:85
8.1888 :8 788:3. 8.8 8:810. 0:. :8 .00: 5 9.8:. row: >30::. 38 8:: 8:8 5 Cm :0:08:. 0:: 80:8: 8 95:3. :.:: 898
0: :8 <98 0:. 08 z. :8 :05. :m 888:. 83 0:82:83. 00830858 :: 00858 2:.8: 0:: 53.25: .8 8:58: 3. 02:85. 88:
3.588 army :.08 8 875.8 8:-8: 0:: 2:15: 3. 505:. 0:88:00: :8 0 95:3. :.:. 8:8 0: :8 <58 0:. 0:8 2. :8 :5. :m
83:88: 8 m 03860:. 8: 88 0:82.88. 4.8... 0:880— ; a 95:3. :.:: 8:8 0: :8 <58 0:, 08 a. :8 :.5: 8 8.288: 8 w 8: :8
8880:. 2.: 83 0:82.88. :83 2.880. :m 0 95:3. :.:. 8:8... 0: :8 <98 0.q 0:8 2. :8 :2: :8 08:98: 8 m :0:-8: :8 078:0:
0:: 83 0:828... 9.880: 83:50: :8 88:8: 00008::m 8 :8 W080 m5 H89... 00.5388: OCH—.0: 0:: :80 853:. ”2.0:: :m 0
95:3. :.2 8:88 0: :8 <58 0:. 0:8 2. :8 88: :08 0:. :8 _00: :8 0:2 33. 0:: 83 0:828... 8:88.88: 8 :8 8855.: 0:. :8
:2: 88: 0808 888:8: :: 8085.8 0:. m0::. W08»: «<0:. 8888 :m :8 8:0 0:. 88: 58:58 8 88— 8888. 800582.08:
88.8 :m :8 8:0 0:. .8. 500:8 8 88: 88.8. romAmma: :m :8 881:5. 0:. :8 :.:..con 0:. :88 8508 :8 503:0: 0:. :8 :5: H0888
8882.08: 888 :m :8 8:0 0:. 8.888 88:. 8 88: 888. £258 005.: :8 :8 98858 83.8: 088:. 88:. .5: 9:85
59::8. romcom: 802.13.: :8 :8 _0mm..:::: 0:. :8 5089:. 0:. :8 :00: m: :38. homes: 05005: 8 :8 _0m0:::: 0:. :8 _00: «:8 ::
Cm 8:08. 8008: :m 0 9.5.3. :.:: 8:8 0: :8 <58 0». 0:8 mm :8 :00: :m 8008:. .5: 88 0:82:88. 25:88 8 “88:888.. 8:08
808: 55:8: 8:68 07888: a: :8 :5. 8.8—. 5...... .2... 0:: .. 8:888 888:8—J. 88:58:: 0. :8 :..\o. :.:... 5: 8.x. 8.8..

 

 

 

7855. how: >805. 80.8: . 788:0. r0»: >500... . 8008:
a: Q: a: 8: G: 3: Q: :.: G: C 0: c .: c N:
9.8.00: 505.... 35:8... .. .. 5. :5 5.5: 5.5: 5.50
8. _ w: 8. :N: 8. 3: 8. :5: 8.8: 8.8:
0:880. .. ”3.05. 5.3.... 5.8.3.... 5.8 5.5 5.5: 5.5:
8. I: 8. 5: 8:8: 8.-: 885: 8.8:
H8..— 05862 50:: 5.8.... 5.5— 5.5: 5.3 5.5
8.3: 8.5.: 8.3: 8.3: 885: 8.3:
.281 2.880: .. :885. 5.8 5.3 5.8 5...: 5.3 5.5:
8.3: 8.3: 8N3 8b.: 8.8: 8.3:
4.8% 2.880. 5.8.....:.5.3..:.:.. 5.8 5.8 5.55 5.5:
8.3: 8.3: 8.5: 8.8: 8.3: 8.8:
H83 038:2 .. ”880...: 5.3:... 5.8:... 5.: 5.5: 5.5: 5.5:

8.3: 8.3: 8km: 8.3: 8.3: 8.3:

78

romawmmoav
r38»?

58302.08. 888

romammav

8.8.
8.88
8.8:
8.5
8.88:
8.8 c

.3555 8882.08— 8828

€2.35 88:22.08— 888

<9: 955:8
H.385. 955:8
0894308
wan—:83

<8
<8
Sum
o. 3

Haw—o r5 r2:— nrsnuanoluanu 2:— Gama—5.23:3

cbc PNNEI e.g.,: -o.~o....:.-o.~a..:.i -cb— Poo 95:... 9N1:
:53 8.8V 8.93 8.8V 8.3V 8.8V 3.8V 8.8V Aobuv
-o.~m -991... .93 93...... :5: 5.8%.... .93 -oho: -93
8N3 8&3 8N8 8.88 3.53 8.3: 8N3 3.N3 Show

8.8:. 8.81.8.3... 8.8 8.88 82:83: 8.8.2.888...
888 8.85 8.3 8.83 8.83 88: 8.88 8.85 8.5
8.8 8.8.... 8.88 8.8 8.5...
883 8.83 8.§ 8.88 8.83
8.8.: 88...: r8... 8.8:. 8.88:...
8.5 8.88 8.88 8.5 8.8:
8.2: 8.88 8.8 8.8.. 8.8
8.28 8.89 8.8: 8.5 8.88

<8 <8 <8 <8 <8 <8 <8 <8 <8
<8 <8 <8 <8 <8 <8 <8 <8 <8
33 G“: 32 Sum 33 Sum Gum 3.: 3w—
o._m Pug 9.3 - .. P: c. 3 com 93

88:82:.
8.83 88$
8.8: :8:
8...“: 8.3
8.3 8.88
8.83 8.8:

o.oa
8. _3
#3....»
808
cow
Bu 3
<8 <8
<8 <8
33 33

79

.550 F5 >255”! 5.1058 0:.— 905.08

 

$3..qu 682.830:

505.00

 

F 3.2: 9.0308138

>m0 008 $0.: 50 :38 8828:0002 8 90 020 0». 0:50:30: 0.. :00 .00:

w0=0<<0mm 88. 0808 0: 80 0:0 08 80 0.008 0108 8 90 $550030:
2.830803

H20— 0808

088 80.8 w0=0£018 @808 80.8 0. 90 0:0 0». 50 0.008 010—. 8 =00 $590030: 0:050:80:

w0=0¢<018 0:35:80 005.080 5:888”. 8x. 0008005030: 0:0 0803820: 0: 50

mm:-0> 0:0 0% 80 0008 0100.8 80 $550010: 03.850302

H00: 8:80.. 0*. 080.868 05. 80 002928 B 80 0:0 08 80 «03 0108 8 80

20:88 05. 0523.08 . .
$590000: 2.850502

5&0 om 8550 0808 8 88. 8805 0: :00 0:0 0m 90 V08 0108 8 90

.5558 8 88— 8808 . .
$590000: 02050300.

8540:: 8808 - 0.50:» =0558v \ H03: 0808 m: 80 0:0 0». :00 «08 010.. 8

Iron—aim 0038. 8 88— 0808 . .
90 $590020: 30:80:03

H03: 000: \ H08— 8808 0» 80 0:0 0m 80 «02. 010.. 8 80 $590020:

r0<08m0 8030
02850303

20. 50030 8088055 00300835 5.00:3 \ H90— 0808 0: :00 0:0 0.. :00 0.08

50030 8 8808 . . .
0:08 8 :00 $590000: 0:050:80:

N. 6040038 000030300

020085: 00:00:88: H08. 0005‘ 88.80 00. 80 80 8: 80:00:01 838030—0088

0038— «0: H08— 0005. 88.80 om 90 80 :0: 8080:0508 858 ..0¢<:0a:8 0800:8080:

.w. 9.0002 60000000138

 

000.800: 0:0
55-500

35-5000

55.58

W5-5:0

55-500

55-58

55-58

35-58

55.500

55-58

55-500

80

4.08. 00008

r0<08m0 8:0

500.00 00:08:00

08.. :02 00:08:00

50080908. 0808

00080.0 08082.08. 0088

H.008 :.:u 3:80:00 (00100.00 00: 5008010003
5.0: 0m 8:0. 0.808 0:. :8 :08 03:08: 0.00 :8 8.00 000000. 0: :8 00: 0:.
:8 0.00.. 010.. 8 :8 85:80:00 02080.80:
508 0:. 8:0. .8558... 8 0:.: 0:. 8:0. 0808 0:. :8 :08 :00: 00.0:m 8 :8 8.00
000000. 0: :8 0:: 0:. :8 0.00: 0:0: 8 :8 08:80:00 0:080:80:

00:08:00 002.00: 0 :88 500083 \ 88. 000089-: 8:0 00: .8

00000020 .8283 \ 8:0. 888?: 8:0 88.8.8 :8 :0: :0 0:00:05.
00:88:: 880 :.0.. :8 8:8 08 08:00.0 $08 :03 .0 55-500 00: :.8
:8 00:08:00 8 00.00.08: 08.. :8 350-800 0010:

0000.08: 002.00: 0 :38 00.0 0:. 800.003 00: 8:808:83 \ 8.0.
888:7: 8:0. 00: .8 000000.... 0.0: 0:. 800083 00: 8:808:83 \ 8.0.
888:7: 8:0 88.8.05 :8 :8 :0 0:00:08. 0080.80: 0080 :.0: :.0 8:00
0.0 08:00.0 :.08 .05.. :0 35-500 00: :.8 :8 0000.080 0.. 00.00.08: 08..

5:8 0:. 800.003 8 00.: 0:. 8:0. 08089-3 0:. :.0 :08 :.0: 00.00m 8 :8 00.00
000000. 0: :00 00: 0:. :8 0.00: 0:00 8 :8 «0.8.03.0: 00880800:

5.00 0:. 88.0.0 0,808 8 8:0. 0:. 88. 8000.. 0:. :8 :08 :8. 00.0:m 8 :8
80.0 000000. 0. :8 0:: 0:. :8 0.000 .010: 8 :8 08:80:00 0.880800:

.3 00:0,... 0:: 83.80.08. 0.003. 003.00.00.00

me... 0005. :.::

.0388: 000.0. :8.

00.0.00. 00:80:00

00:0. 0008 0:. 000.0. 00.: 00. 80.0.4.8 000.8

00:0. :88 0:. 09:0. 00.: 0v. 8:8:000. 5.8008 88.098 8.088 000.6.
00083

00.0.00. 8100.0 :8: 8.08 00 :8 8.8 0:. 0:0 .:. 0:0 0:. :8 :0:.m 80800.88
A0003. 0800808 V1.0: 0.8 0 08.000: 0:. :8 20:000. .8880? 0:. .880 :0
:8 0.000 .0 £080 :.0 .000 $8 80:80:09 .288 «<08 0 :08 Q: 000.: :.0:
£08 00:80.0. 0000008: .0 :8 0006.0 00: :8. $08 000000. 5884.00?
:88. 0:: £00.80. 300$. 3:85.08. 0:.:=08:

.

 

85-500

55-500

55-500

55-500

55-500

W5...

.25...

.05.. 00:
22288002808050...

81

qua—o r; >355?" sins—nu E:— moEdam A255
m. Mcaﬁnﬁnm has. Granaanawaa

be»: 852:: row: ”59:: a: Cw 8:83

 

Unwrap:
Z353! row: Buns—.5. a: mwwmv Davao»:
moor—Ba 95.3% :5" 58m 2. En <35». 3. 25 =. Ea 58.. mm moo—=3. No.6 oﬁnnﬁmmo Una—mow:
a. 931:: Enauﬁnaae:
Qing— 95:5~ :54 grow o: :5 $55 cm on» :.:.» an: a awn—58a 8 a 2525—. 53.0 mpg 53383
Nmno Bro—25mm
ﬁn: 3530— 95:3‘ :54 gram 0: En <32... cacao _mEo an: a warm—8a 8 a 8H. m 0..—mag... E0 23 55383
N03 Brogan
.22.». £532 USESV 9E grow on En.<m_=o on 28 _m 90 m2: 5 warm—8a 8 w arm—ova En Ea 53383
Saran 0.8. ~03 0902250
.. Um.» $.03 2.01am; cw Em Q98» Ema—.838 Monica mam.

 

82

1. 9 BIBLIOGRAPHY

[1] Allen, F. and D. Gale, 2000. “Comparing ﬁnancial systems,” MIT Press,
Cambridge, Massachusetts.

[2] Anderson, R. C., Mansib, S. A., Reeb, D. M., 2003. Founding family
ownership and the agency cost of debt. Journal of Financial Economics 68,
263—285.

[3] Bae, K.-H., Kang, J .-K., Kim, J .-M., 2002. Tunneling or value added?
Evidence from merger by Korean business groups. Journal of Finance 57,
2695—2740.

[4] Baek, J .-S., Kang, J .-K., Park, K. S., 2004. Corporate governance and ﬁrm
value: Evidence from the Korean ﬁnancial crisis. Journal of Financial Economics
71, 265-313.

[5] Bank for International Settlements. Securities statistics and syndicated loans.
11 t tp z/fw w w. bis_.org/s t. atistics/sccs tatsJitm.

[6] Beck,T., Levine, R., Loayza, N., 2000. Finance and the sources of growth.
Journal of Financial Economics 58, 261-300.

[7] Bhojraj, S., Sengupta, P., 2003. Effect of corporate governance on bond
ratings and yields: The role of institutional investors and outside directors.
Journal of Business 76, 455-475.

[8] Bosch, 0., and Stefen, S., 2008. Informed lending and the structure of loan
syndicates — Evidence from the European syndicated loan market. Working paper.

[9] Carletti, E., 2004. “The structure of bank relationships, endogenous
monitoring and loan rates,” Journal of Financial Intermediation 13, 58-86.

[10] Carletti, E., V. Cerasi, and S. Daltung, 2007. “Multiple-bank lending:
Diversiﬁcation and free-riding in monitoring,” Journal of Financial
Intermediation 16, 425-451.

[11] Chang, J.J., Shin, H.-H., 2007. Family ownership and performance in Korean
conglomerates. Paciﬁc-Basin Finance Journal 15, 329—352.

[12] Chang, S.-J., 2003. Financial crisis and transformation of Korean business
groups. Cambridge University Press, Cambridge, United Kingdom.

[13] Chiu, M. M. and Joh, S. W., 2004. Loans to distressed ﬁrms: Political
connections, related lending, business group afﬁliation and bank governance.
Working Paper.

[14] Choi, J. P. and Cowing, T. G. 1999. Firm behavior and group afﬁliation: The
strategic role of corporate grouping for Korean 'rms. Journal of Asian Economics
10, 195-209.

83

[15] Claessens, S., S. Djankov, J. P. H. Fan, and L. H. P. Lang, 2002.
Disentangling the incentive and entrenchment effects of large shareholdings.
Journal of Finance 57, 2741-2771.

[16] Demirgiic-Kunt, A., and Maksimovic, V., 1998. Law, ﬁnance, and ﬁrm
growth. Journal of Finance, 53, 2107-2139.

[17] Detragiache, E., T. Tressel, and P. Gupta, 2008. Foreign banks in poor
countries: Theory and evidence. Journal of Finance, forthcoming.

[18] Diamond, D., 2004. Committing to commit: Short-term debt when
enforcement Is costly. Journal of Finance 59, 1447-1479.

[19] Esty, B. C., Megginson, W. L., 2003. Creditor rights, enforcement, and debt
ownership structure: Evidence from the global syndicated loan market. Journal of
Financial and Quantitative Analysis 38, 37-59.

[20] Faccio, M., R. W. Masulis, and J. J. McConnell, 2006. Political connections
and corporate bailouts. Journal of Finance 61, 2597-2635.

[21] Gatev, E., Strahan, P., 2008. Liquidity risk and syndicate structure. NBER
Working Paper 13802.

[22] Ghemawat, P., Khanna, T. 1998. The nature of diversiﬁed business groups:
A research design and two case studies. Journal of Industrial Economics 46,
35—61.

[23] Giannetti, M., 2003. Do better institutions mitigate agency problems?
Evidence from corporate ﬁnance choices. Journal of Financial and Quantitative
Analysis 38, 185-212.

[24] Giannetti, M., and S. Ongena, 2007. Financial integration and ﬁrm
performance: Evidence from foreign bank entry in emerging markets. Review of
Finance, forthcoming.

[25 Harman, T. H., 1997. Market share inequality, the number of competitors,
an the HHI: An examination of bank pricing. Review of Industrial Organization
12, 23-35.

[26] Holmstrom, B., and Tirole, J ., 1997. Financial intermediation, loanable funds,
and the real sector. Quarterly Journal of Economics 112, 663—691.

[27] Hoshi, T., A. Kashyap, and D. Scharfstein, 1990. The role of banks in
reducing the costs of ﬁnancial distress in Japan. Journal of Financial Economics
27, 67—88.

[28] Hwang, L.-S., K. Park, and R. Park, 2008. Corporate governance and payout
policy: Evidence from Korean business groups. Working paper.

[29] Ivashina,V., 2009. Asymmetric information effects on loan spreads. Journal
of Financial Economics, 92, 300-319.

[30] Jang, H-S Corporate governance and economic development: The Korean

84

Experience in Democracy Market Economics and Development

[31] Jensen, M., Meckling, W., 1976. Theory of the ﬁrm: Managerial behavior,
agency costs and ownership structure. Journal of Financial Economics 3, 305—360.

[32] Joh, S. W., 2003. Corporate governance and ﬁrm proﬁtability: Evidence from
Korea before the economic crisis. Journal of Financial Economics 68, 287-322.

[33] Jung, K., B.-M. Kim, and Y.-C. Kim (2001). Governance discount: Valuation
of the Korean business groups. Working paper.

[34] Kroszner, R., 1998. The political economy of banking and ﬁnancial
regulatory reform in emerging markets. Research in Financial Services 10, 33-51.

[35] Krueger, A. O. and J. Yoo, 2002. Chaebol capitalism and the currency
ﬁnancial crisis in Korea. In S. Edwards and J. A. Frankel, eds.: Preventing
currency crises in emerging markets, University of Chicago Press, 601 - 662.

[36] La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R. W., 1997. Legal
determinants of external ﬁnance. Journal of Finance 52, 1131-1150.

[37] La Porta, R., Lopez-de—Silanes, F., Shleifer, A., Vishny, R. W., 1998. Law
an ﬁnance. Journal of Political Economy 106, 1113-1150.

[38] Lee, Y.-H., 2005. Financial structure and ﬁnancing of chaebol (in Korean).
N anam Publishing House. '

[39] Levine, R., 1999. Law, ﬁnance, and economic growth. Journal of Financial
Intermediation 8, 8-35.

[40] Lim, W., Haggard, S., Kim, E., 2003. Economic crisis and corporate
restructuring in Korea. Cambridge University Press, Cambridge, United Kingdom.

[41] Min, B. S., 2007. Changing pattern of corporate governance and ﬁnancing in
the Korean chaebols. Economic Papers - Economic Society of Australia 26, 211-
230.

[42] Nam, C., Kang, Y. and Kim, J .-K., 1999. Comparative corporate governance
trends in Asia. In: Corporate governance in Asia: A comparative perspective.
OECD, Paris.

[43] OECD, 2000. Regulatory reform in Korea. OECD Review of Regulatory
Re orm, Paris.

[44] Petersen, M. A. and R. G. Rajan, 1994. The beneﬁts of lending relationships:
Evidence from small business data. Journal of Finance 49, 3-37.

[45] Qian, J ., Strahan, P., 2007. How laws and institutions shape ﬁnancial
contracts: The case of bank loans. Journal of Finance 62, 2803-2834.

[46] Rajan, R., Zingales, L., 1998. Financial dependence and growth. American
Economic Review 88, 559—586.

[47] Rajan, R., Zingales, L., 2003. The great reversals: The politics of ﬁnancial

85

development in the 20th century. Journal of Financial Economics 69, 5-50.

[48] Schneider, M. and A. Tornell, 2004. Balance sheet effects, bailout guarantees
and ﬁnancial crises. Review of Economic Studies, 71, 883-913.

[49] Shin, J.-S. and Chang, H.-J., 2003. Restructuring Korea Inc: Financial crisis,
corporate reform, and institutional transition. Routledge.

[50] Shin, H.-H., Park, Y. S., 1999. Financing constraints and internal capital
markets: Evidence from Korean ‘chaebols’. Journal of Corporate Finance 5,
169-191.

[51] Shleifer, A. and Vishny, R., 1992. Liquidation values and debt capacity: a
market equilibrium approach, Journal of Finance 47, 1343-1366.

[52] Shleifer, A., Vishny, R., 1997. A survey of corporate governance. Journal of
Finance 52, 737—783.

[53] Stein, J., 2003. Agency, information, and corporate investment. In:
Constantinides, G., Harris, M., Stulz, R. (eds)., Handbook of the Economics of
Finance. Elsevier, North Holland,

Amsterdam.

[‘54] Su', A., 2007. Information asymmetry and ﬁnancing arrangements: Evidence
om syndicated loans. Journal of Finance 62, 629-668.

[55] Tirole, J ., 2005. The theory of corporate ﬁnance.

[56] Vives, X, 2006. Banking and regulation in emerging markets: The role of
external discipline. The World Bank Research Observer 21, 179-206.

[57] Wurgler, J ., 2000. Financial markets and the allocation of capital. Journal of
Financial Economics 58, 187-214.

[58] Youngmo, Y., 1999. Chaebol reform: the missing agenda in “corporate
governance”. In: Corporate governance in Asia: a comparative perspective.
OECD, Paris.

86

CHAPTER 2

Creditors’ Rights and Loan

Structure

Abstract

In loan syndicates, the participant lenders, who are aware of the conflict of interest
between shareholders and creditors, are hesitant to participate unless they are sure
that the lead arranger(s) increases ex post monitoring efforts to deter the ﬁrms from
acting for the interest of shareholders at the expense of the creditors’ beneﬁts. How-
ever when the ﬁrms strengthen the protection of the creditors’ rights, the intensity of
monitoring required by the participant lenders will be lightened, which is reﬂected in
the structure of syndicated loans by smaller loan concentration or lower loan share
held by the lead arrangers. Using proxies (the G index, state antitakeover law, and
state payout restriction) for the degree of the protection of creditors’ rights of the
ﬁrm, I observe that the average loan concentration of the ﬁrms, which set up more
provisions for protecting creditor’s rights, is lower and the likelihood of the creditors’
holding collateral is smaller. However the extent of the impact of such protection on
the loan concentration also depends on the degree of the product market competition

the firms face. For the ﬁrms that belong to the highly competitive product markets,
87

I found that the impact of the G index is pronounced and the creditors are less likely
to hold collateral, and embed fewer ﬁnancial covenants when the ﬁrms are prone to

protect creditors’ rights.

2. 1 Introduction

Conﬂicts between shareholders and creditors can arise because of different payoff
structures of equity and debt (Jensen and Meckling, 1976). Shareholders claim the
upside payoffs of the ﬁrm whereas creditors have to bear the downside risks. To guard
themselves from shareholders’ risk-shifting(asset substitution) incentives, creditors
may articulate loan contracts by adding covenants, or holding the ﬁrm’s assets as
collateral. Apart from these restrictions, borrowers are required to pay larger interest
when creditors believe they will be exposed to, with higher probability, the chance of
being involved in conﬂicts in the future (see e.g., Chava, Livdan, and Purnanandam
(2009)).

The conﬂicts affect not only the terms of loan contracts, but also the structure of
loans, such as the shares of the loan held by the lead arranger(s) in the syndicated
loans which are one of the major vehicles of corporate ﬁnancing for US ﬁrms. As
emphasized in Focarelli, Pozzolo, and Casolaro (2007), and Suﬁ (2007), the shares of
the loan held by the lead arrangers is of importance, because the shares translates
into the intensity of monitoring by the lead arrangers on the borrowers. When the
ﬁrm’s payoffs depend on the degree of creditors’ monitoring efforts, as the loan share
held by the lead arrangers gets larger, payoffs are also expected to be greaterl.

This paper investigates the link between shareholder-creditor conﬂict and its im-

 

1Focarelli, Pozzolo, and Casolaro (2007) observe that announcement effect of the new loan on
the stock price is an increasing function of the portions of the loan retained by the arranger be-
cause the shareholders expect that the lead arrangers as delegated monitoring agents will increase
their monitoring efforts when their shares of loan increase. Chen, Harford, and Li (2007) found
that concentrated share holdings by independent long-term institutions are related to post-merger
performance.

88

pact on the structure of syndicated loans with the question of whether the creditors
are worried about the risk-shifting problem and then how the ﬁrm’s and the state’s
measures of creditor rights protection affect the structure of syndicated loans using
US syndicated loans, and the ﬁrm’s antitakeover provisions, and the state’s takeover
law and payout restriction. Furthermore, under which environment, the impact of
those protection measures on the structure of syndicated loans becomes greater.
The sample data I use in the paper are drawn from the Dealscan database dis-
tributed by the Loan Pricing Corporation for US ﬁrm loans from January of 1990 to
June of 2007. and the borrowers are publicly traded ﬁrms. The beneﬁt from using
publicly traded ﬁrms is to exclude, in some limited extent, the effects of asymmetric
information problem between the lead arrangers and the participant lenders on the
structure of syndicated loans 2
In the previous studies on the syndicate loan structure (see, Dennis and Mullineaux
(2000), Ivashina (2009), Jones, Lang, and Nigro (2005), Lee and Mullineaux (2004),
Suﬁ (2007)) 3, agency problems stemming from adverse selection and moral hazard
are found to affect the loan share distribution between the lead arrangers and the
participant lenders, as well as the loan spread over Libor. Adverse selection problem
occurs when the information on the borrower is not fairly revealed to all lenders, such
information inequality can arouse suspicion that the lead arrangers may be selling a
lemon project to other lenders and as a consequence interested lenders may hesitate
to join the syndication. On the other hand, after contracting, the participant lenders
face moral hazard problem caused by the lead arrangers. Especially, the participant

lenders want to make it sure that the loan which they lend does not go bankrupt,

 

2Although the ﬁrm is publicly traded, some ‘soft’ information is not believed to be shared by all
lenders. However publicly traded companies can be considered to cause less information asymmetry
problem than private ﬁrms (Suﬁ, 2007). With public firms, the effects of the degree of creditor
rights protection is more fairly singled out by excluding the unobserved effects in part stemming
from asymmetric information problem.

3These papers focused on the effects of information opaqueness on the structure of syndicated
loan.

89

or its repayment schedule is not renegotiated in the future. Without the certainty of
the borrower being monitored by the lead arrangers, interested lenders also are not
likely to participate in the loan syndication. To ameliorate the concern of the ex: post
monitoring problem, the lead arrangers take more shares of the loan. However, when
the lead arrangers are reputed in the previous years for their commitment in paying
due diligence on the borrowers, the concern of the ea: post monitoring problem will
be mitigated.

In addition to the reputation effects, the degree of creditor rights protection that the
ﬁrm has installed can affect the structure of syndicated loans. When the management
of a ﬁrm is more aligned with shareholders’ rights than with creditors’, interested
lenders will hesitate to participate in a syndication of the loan because of a lower
possibility of their rights being honored against shareholders’ rights when the ﬁrm
falls into a situation where the two rights are in conﬂict. In other words, for ﬁrms that
employ good measures of protection, the participant lenders, due to lessened concern
about conﬂict, will demand less monitoring on whether their rights are compromised.
This will result in smaller shares of a loan held by the lead arrangers. This aspect is
not dealt with so far except for. the analysis on such effects on loan pricing(Chava,
Livdan, and Purnanandam, 2009)

To measure the degree of ﬁrm’s creditor rights protection, I use the G index and
E index (see details in the next section) as in Gompers, Ishii, and Metrick (2003)
and Bebchuk, Cohen, and Ferrell (2009). In addition to the ﬁrm-level protection
measures, there are state-level protections: the states employ the payout restriction
(Wald and Long, 2007) and antitakeover laws, both of which vary across the states
in terms of their severity.

The ﬁrst ﬁnding is that ﬁrms which are more protective of creditors’ rights have
their loan less concentrated in terms of the share of the loan held by the lead ar-

rangers, and constructed loan concentration index, Herﬁndahl-Hirschman (henceforth

90

Herfindahl), while controlling firm and loan characteristics. As the firms weaken their
shareholder’s rights (it amounts to the increase in the G and E indices) by increasing
one unit of the G index (max =24), the loan concentration decreases on average by
2%. The decline is more pronounced when the ﬁrms are Dictatorship type ones (very
repressive on shareholder’s rights); the loan concentration is smaller on the average by
11.4% compared with that of the other control group. Furthermore, I estimated the
size of the effects of state level protection(proxied by antitakeover provisions), con-
structed by grouping states according to how restrictive the state is toward takeover,
based on the number of ﬁve antitakeover provisions used in Bebchuk and Cohen
(2003). To my surprise, the state law contributes to the decline of loan concentra-
tion, in terms of shares held by the lead arranger, by about 18.5% regardless of the
degree of the severity of the protection,(“moderate”, or “most restrictive”). This ﬁg-
ure is about 10(18.5%/ 2%) times larger than the case when I use the G index index
instead. This ﬁnding demonstrates that the participant lenders in the loan syndica-
tion look at whether the borrowing ﬁrms provide more protection for creditors and
furthermore to what extent they do.

Creditors’ rights are also exposed to the risk of being infringed when the ﬁrm’s
manager compensation scheme is aligned with the interest of its shareholders (see,
Vasvari (2009)). This paper observes that although the manager’s compensation is
more dependent on the change in stock price and its volatility, which translates into
that where the manager has more incentive to increase shareholders’ wealth, the
effects of the protection of creditors’ rights is preserved in tact.

To advance the investigation on the role of institutions related to the structure
of syndicated loans, I examine the relationship between the state’s level payout re-
striction and loan structure (whether the loans are “secured”). Payout restriction is
a state law which prevents the ﬁrms from paying out dividends unless they satisfy

the threshold level of capital sufficiency. I ﬁnd that the payout restriction strongly
91

affects the decision of whether the creditors hold the ﬁrms’ assets as collateral in their
syndicated loan contracts. The holding of collateral is also found to be affected by
the degree of the protection of creditor’s rights, identiﬁed by the G index, and the E
index.

I also examine the effectiveness of the protection of creditor’s rights in the environ-
ment, in which creditor’s rights can be undermined, such as when the ﬁrms belong
to competitive industries (Giroud and Mueller, 2010). The eﬂectiveness of the pro-
tection of creditor’s rights is shown more greatly when the industries to which the
borrowing ﬁrms belong are very competitive in their product markets. Competitive
pressure will enforce discipline on managers (Giroud and Mueller, 2010) in such a way
that the managers pay more cash to their shareholders (Grullon and Michaely, 2005).
For such reason, the creditors become more attentive to ﬁrms’ protective measures
for creditor’s rights when the ﬁrms belong to highly competitive industries.

This paper also contributes to the literature on the eﬂects of institutional differences
on the structure of syndicated loans (e.g., Bae and Goyal (2009), Qian and Strahan
(2007)) 4. Different from cross-country studies where differences in countries’ creditor
rights are one of major explanatory variables in ﬁnding the determinants of the loan
structure, the paper demonstrates that even under the same national jurisdiction,
legal and institutional differences across states, (mainly in the state’s antitakeover
provisions, and in payout restriction) affect the shaping of loan structure.

The remainder of this paper is organized as follows, section 2 will discuss the
role of creditor rights protection on loan contracts, and the environments where the
impact of creditor rights protection can be affected. Section 3 explains data and
major statistics. Section 4 discusses empirical speciﬁcation and key variables that are

included in the regressions. Section 5 presents main empirical results and section 6

 

4Bae and Goyal (2009) found that lenders reduce loan amounts, shorten loan maturities and
increase loan spreads when the enforceability of contracts is believed poor. Qian and Strahan (2007)
observe that under strong creditor rights protection, loans have more concentrated ownership, longer
maturities, and lower interest rates 92

discusses endogeneity, and the ﬁnal section ends with the conclusion.

2.2 Background

In this section, I will explore how protection of creditors’ rights can affect the structure
of syndicated loan. Furthermore, The exploration is extended to the question that
in which environment the protection of creditors’ rights are perceived effective to the

creditors and the consequent implications on the structure of loans.

2.2.1 Protection of Creditors’ Rights and Structure of Loans

A syndicated loan is a type of loan which is funded jointly by more than two lenders.
The lead arranger(s), which can be a sole lender, or multiple lenders, establishes a
relationship with the borrower and determines the terms of the loan contract. The
lead arranger then seeks other lenders who can fund part of the loan. The lead
arranger can have more information on the borrower through the relationship than the
participant lenders. Such information asymmetry in turn, can affect the participation
decision of the non-lead lenders. The non-lead lenders are reluctant to participate
unless they are certain that the lead arrangers will do what they have to do, as the
lead arrangers, for them, not selling a lemon, and monitoring the borrower after the
deal is struck. If the participant lenders do not join the loan syndication, which
results in the lead arrangers taking a larger portion of the loan instead.

Previous studies found that the asymmetric information problem which can cause
mistrust affects the structure of syndicated loans. When the borrower is an opaque
private ﬁrm, which entails an information asymmetry problem, the lead arrangers
have to increase their funding shares of the lean more than when the borrower is a
public ﬁrm which is relatively more transparent (Suﬁ, 2007). The lead arrangers who

take more loan shares will be exposed to the possibility of a bigger loss when the

93

borrower defaults. To avoid such a risk, the lead arrangers make efforts to provide
more information on the borrower to the participating lenders, or build a reputation as
credible lead arrangers by paying due diligence on the loans which they have arranged
in the past 5.

N 0t only the lead arrangers’ efforts, but also the legal and institutional environment
in which the ﬁrms do business, can affect the structure of syndicated loans. Depending
on how securely creditor rights are protected, the number of the lenders, and the
shares of the loan held by the lead arrangers are determined. Qian and Strahan (2007)
found by using country cross-sectional syndicated loan data that in the country where
creditor rights are more strongly protected, loans have longer maturities, and lower
interest rates. Their ﬁndings are also attested in the recent study by Bae and Goyal
(2009) in which they also observe that the lenders reduce loan amounts, shorten loan
maturities, and increase loan spreads when the loans were initiated in the environment
which provides poor enforceability of loan contracts.

In’line with’the previous studies, I conjecture that the structure of syndicated
loans is also affected by creditors’ concern about the ﬁnancial risks that the creditors
will face when the borrowing ﬁrm acts for the beneﬁts of its shareholders at the
expense of the creditors’ interest. favoring shareholders can make the conﬂict between
shareholders and creditors complicated, such as by shifting risks from the one to the
other, or by paying excess payout. In the loan initiation stage, the concern about
being involved in conﬂict in the future can hamper the participation of participant

lendersf’.

 

5Gopalan, Nanda, Yerramilli, and Drive (2009) found when the ﬁrm for which a bank arranged
a syndicated loan went bankrupt, the lead arranger suffered from the drop of aggregate level of
activity in the loan syndication market in the following years.

6In relation with takeover concern for creditors, some studies explored the impact of takeover
concern on the spread; Chava, Livdan, and Purnanandam (2009) found that when the borrowing
firm is more likely to participate in takeover, loan interests are charged higher to compensate risk
that can be increased by takeover. The relation between takeover and structure of securities also
has been studied. Klock, Mansi, and Maxwell (2005) observed that anti-takeover provisions lower
the cost of debt ﬁnancing. Cremers, Nair, and Wei (2004) also found that shareholder control is

94

For the measure of the degree of creditor rights protection, I use the G index in
this paper which was constructed by Gompers, Ishii, and Metrick (2003). This index
quantiﬁes how much a ﬁrm is protected from hostile takeover threats. The ﬁrm takes
various antitakeover measures to protect itself from hostile takeover offers by adding
takeover defenses and some restrictions on shareholder rights, such as staggered terms
of directors and severance packages for managers. Along with the individual ﬁrm’s
efforts, many states also join in the antitakeover movement by stipulating legal restric-
tions against hostile bids. Such efforts still continue to exist and vary across states.
These antitakeover measures can work against the interest of shareholders. Gompers,
Ishii, and Metrick (2003) found that ﬁrms, whose G index is higher, experience lower
rates of stock return. From the creditors’ point of view, such measures secure their
rights by decreasing the probability of the creditors being involved in the situation
where their seniority is compromised when the ﬁrm is acquired, or merges with other
ﬁrms, resulting in the rise of its ﬁnancial leverage.

Gompers, Ishii, and Metrick (2003) construct the G index using the IRRC, which
consists of ﬁve categories; Delay(tactics for delaying hostile bidders), Voting(voting
rights), Protection(director/officer protection), Other(other takeover defenses), and
State(state laws). The 24 provisions comprise the index. In addition, Bebchuk and
Cohen (2003) also constructed an index for the degree of state-level antitakeover
measures (the index is obtained from Lucian Bebchuk’s website), having a maximum
value of ﬁve when a state has stipulated the complete set of the following laws: ‘Con-
trol Share Acquisition Statute’, ‘Fair Price Statute’, ‘Business Combination Statute’,
‘Poison Pill Endorsement Statute’, and ‘Constituencies Statute’. The index is very
persistent and does cover from 1986. Since it covers only up to the year 2001. I apply
the last year’s index to the uncovered years after 2001.

On the other hand, along with antitakeover laws, many states stipulated the law

 

associated with lower bond yields if the firm is protected from takeovers.

95

which restricts ﬁrm’s payout which is carried either by paying dividends, or by pur-
chasing its shares? Such restriction on shifting ﬁrm’s cash or assets at the expense
of creditors’ rights to shareholders will assure the participant lenders in joining the

syndication of loans.

2.2.2 Product Market Competition and Managerial Incen-
tive

The protection of creditors’ rights can be attenuated in the environment where prod-
uct market competitions are severe. The CEO of a ﬁrm, which belongs to a competi-
tive industry, will be pressured to maximize the value of the ﬁrm. In other words, the
interest of shareholders. In non-competitive industries where lack of competitive pres-
sure fails to enforce discipline on managers, the managers are not very motivated to
act in the interest of shareholders 8. Creditors who perceive the degree of competition
will hesitate to join a syndication when it is too severe.

On the other hand, since 1990, many US ﬁrms started to compensate their managers
with stock and stock options. Core, Guay, and Verrecchia (2003) observe that CEO’s
compensation by stock and stock options comprises 80% of their total compensation.
Such a skew toward stock compensation can concern creditors. The infringement of
creditors’ rights can take place when the managerial incentive structure is designed

in the form of stock and stock options. Since the managerial compensation is aligned

 

7In California, The restrictions on “Distribution to its shareholders (the transfer of cash or
property by a corporation to its shareholders) include the following language: Neither a corporation
nor any of its subsidiaries shall make any distribution to the corporation’s shareholders (Section 166)
except as follows: (a) The distribution may be made if the amount of the retained earnings of the
corporation immediately prior thereto equals or exceeds the amount of the proposed distribution.
(b) The distribution may be made if immediately after giving effect thereto: (1) The sum of the
assets of the corporation (exclusive of goodwill, capitalized research and development expenses and
deferred charges) would be at least equal to 1 1/4 times its liabilities (not including deferred taxes,
deferred income and other deferred credits);...(quoted from Wald and Long (2007))

8Gompers, Ishii, and Metrick (2003) show that ﬁrms with good governance earn signiﬁcantly
higher stock returns.

96

with its stock price change, the CEO is more likely to act in the interest of shareholders
at the expense of creditors’ rights, such as by substituting assets, and taking risky
projects”. In response to such infringements, rational creditors will ask the borrower
in question for a higher loan spread (see, Levine and Hughes (2005), Vasvari (2009)).
In syndicated loans, as the managerial compensation is more aligned with stock price
and volatility, the degree of creditor rights protection will be lessened and there will
be a decline of its impact on loan concentration.

To see whether the impact of the protection of creditors’ rights is affected by the
managerial compensation structure, the degree of how much the CEO is incentivised
to act in the interest of shareholders, delta and vega 1(are computed using the method-
ology as in Core and Guay (2002) and Edmans, Gabaix, and Landier (2008). Delta
is deﬁned as the dollar change in a CEO’s stock and option portfolio for a 1% change
in stock price, and Vega, as the dollar change in a CEO’s option holdings for a 1%

change in stock return volatility.

2.3 Data and Statistics

The data used in the paper consists of three main categories: data on syndicated
loans, data for constructing ﬁrm characteristics, and data on the measures of creditor
rights protection(corporate governance, state-level antitakeover laws, and state-level
payout restriction). The initial loan data set withdrawn from Dealscan has 16,019
deal loans which are matched with the COMPUSTAT data while excluding ﬁnancial

service ﬁrms. In the process of identifying the borrowers with the COMPUSTAT

 

9Empirically, Coles, Daniel, and Naveen (2006) found that higher sensitivity of CEO wealth to
stock volatility (vega) implements riskier policy choices, including relatively more investment in
R&D, less investment in PPE, more focus, and higher leverage.

1Hanlon, Rajgopal, and Shevlin (2003) found that stock option grants to executives are consistent
with an incentive alignment between managers and shareholders. The inclusion of the value of the
option portfolio is also consistent with Gore and Guay (1999) who argues that stock options provide
better incentives than stock holdings

97

data, I use the ticker, the ﬁrm name, and the ﬁrm’s primary SIC provided in the
loan data. I merge then the loans with the G index, and E index, however, because
of the small ﬁrm coverage of the two indices, a sizable number of observations are
lost, leaving 3,316 loans. Some loan deals consist of a couple of facilities diﬂering in
loan amount (tranche) and maturity. Since the loan structure is determined in the
origination of the deal, when there are more than two facilities, I select the one whose
tranche is the largest among them.

The G index is obtained from Andrew Metrick’s website 1.1 The index is available
for the years, 1990, 1993, 1995, 1998, 2000, 2002, 2004, and 2006. For the loans
whose origination years do not match with their corresponding year’s G index, I used
the index whose year is most close to the years of loan origination as Cremers and
Nair (2005) did. During the course of the years, the G index has been very stable.
Furthermore, because of some caveats raised by Bebchuk, Cohen, and Ferrell (2009)
on the use of the G index as a proxy for ﬁrm’s antitakeover measures, I also use the
E index constructed by them. However, because the E index quantiﬁes antitakeover
measures over a small number of antitakeover provisions, the E index does not vary
much across the ﬁrms, so that the inferences of coefﬁcients in the regressions are
somewhat limited. For this reason, the analysis mainly will be carried on with the
G index. Panel A of Table 2.1 reports the composition of the 3316 loans according
to the G index. The observations are more populated around the index, from 7 to
11, skewed to the right in its distribution with a mean of 9.0 (The E index averaged
around 2.4).

Panel B of .Table 2.1 presents loan statistics. The median loan size is 250 million US
dollars with the mean of 518 million dollars. The average and median loan maturity
are both about 3 years. Less than half of the loans were required to have collateral to

secure the lender’s claim on some designated set of the borrower’s assets. The ﬁrms

 

1Thank for generous provision of data

98

enter into loan contracts for various purposes. The most common reason is working
capital/ corporate purposes, at 54% and the second reﬁnancing at 24%, and the third,
acquisitions at 12%. Although Dealscan provides a large number of loans for US.
ﬁrms, the detailed information on the loan shares held by each lender is rare. When
I eliminated the loans that do not contain the loan share information, only 20% of
the loans survived. In measuring loan concentrations, I use the method used by Suﬁ
(2007). Dealscan provides two headings that reference the role of lenders. I consider
the lenders under the heading, “Lead Arranger” as the lead arrangers. Each lead
arranger holds, on average about 23% portion of a loan, and its median is 17.5%. As
an alternative measure for the loan concentration, I construct a Herﬁndahl-Hirschman
index (henceforth Herﬁndahl) using detailed lenders’ loan shares, theoretically which
range up to 10,000. The average Herﬁndahl in the sample is 1948 and its median is
1357.

In addition to the state-level antitakeover laws, the states impose a payout restric-
tion on the ﬁrms which are incorporated under the jurisdiction of the states. The
state-level payout restriction varies across states by the minimum requirement on the
ratio of book capital to debt before a dividend payment or share repurchase can be
made. Since this is the state law, whether or not creditors code it intoloans as a
covenant, all ﬁrms are subject to it. The purpose of the restriction is to curb a moral
hazard of the ﬁrms so that the creditors’ rights of the ﬁrms are not infringed (Wald
and Long, 2007). The data for the payout restriction were obtained from the website
of John Wald (http://faculty.business.utsa.edu/jwald/).

The Standard and Poor’s COMPUSTAT database provides ﬁnancial data for con-
structing ﬁrm’s ﬁnancial characteristics. The median ﬁrm size in terms of total assets
is 1.3 billion dollars and its mean is 4.6 billion dollars, which indicates the sample
ﬁrms are large ones. The large difference between mean and median also presents

the distribution of the ﬁrms in the sample is skewed to the right. The mean leverage

99

ratio stands at 0.3, as its median does. The ratio of EBITDA to total sales is 0.17 as
the mean and 0.14 as the median.

The Dealscan data set provides detailed loan covenant information. The covenants
are grouped into four categories: ﬁnancial covenants, voting covenants, prepayment
covenants, and dividend restriction. Fifteen ﬁnancial covenants have upper or lower
bounds of restriction on the ﬁnancial variables, such as leverage, net worth, and
current ratio. An average loan has 2.5 ﬁnancial covenants and the most popular
covenants are Debt to EBI T DA, and Interest coverage. Prepayment category consists
of ﬁve covenants that require the borrower to repay the loan on the occasions when
the borrower issues equity, or debt, or sells its assets, or has an excess cash ﬂow.
Except for the covenant, ‘insurance proceeds’, the other four covenants are called
‘sweep’. The four covenants articulate the percentage of prepayment of the loan
when the borrower violates those covenants. For example, the covenant, 50% debt
sweep, means that when the ﬁrm does not honor the covenant, it is required to repay
50% of the loan. Asset and debt issuance sweeps are the most popular among the
sweep covenants in the sample.

The voting category has three restrictions regarding voting on the change in the
terms of a loan. ‘Term changes’ restricts the minimum voting shares in modifying the
life of the loan, requiring 98% approval on average in the sample. The other remain-
ing covenants, ‘Required lenders’ and ‘Collateral release’ are used for non-material
amendments and release of a lien on collateral. Lastly, the dividend covenant puts a
restriction on the activity of cash distribution to its shareholders. This restriction is
coded by a binary variable, ‘Yes’ or ‘No’, indicating whether the covenant is included.

To construct the delta and the vega, I merge the Compustat data and the Execu-
Comp data at the end of every ﬁscal year. Incentives comes from the CEO’S holdings

of shares and options 1.2 CEO’s delta and vega are 1057 and 141 thousand dollars on

 

lThe delta and the vega are calculated based on the Black-Scholes formula,

100

average, and 193 and 38 thousand dollars on median respectively. Market competi-
tion is the Herﬁndahl—Hirschman index (henceforth HHI) that is deﬁned as the sum

of squared market shares in terms of the ﬁrm’s sale volume13

2.4 Empirical Speciﬁcation and Key Variables

I test the cross-sectional relationship between the degree of creditor rights protection

and the loan concentrations. The primary model is speciﬁed as follows:
Loan_Con,;]-t = a+7t +ﬂProtection of Creditor’s Rightsjt +th6 +Xit§ +523, (2.1)

where Loan_Con,jt indicates either Share Held by Lead, or Herﬁndahl. X jt rep-
resents ﬁrm characteristics, Xz-t loan characteristics, and ’Yt year dummy variables.
The linearity model speciﬁcation between the dependents and the variable, Protection
of Creditor’s Rights is based on the observations on the relationship between them.
Figure 1 presents the linearity between the two loan concentrations and the G index
using a non-parametric regression, Lowess.

I use a couple of ﬁrm-speciﬁc characteristics to control for their impact on the
dependent variable. For a measure of ﬁrm size in the literature, I used the loga-
rithm of the ﬁrm’s total assets (item AT in COMPUSTAT) to control the ﬁrm size
effects such as higher reputation effects of the borrower, and the lower default risk
perceived in the market as the size gets larger. To control more directly the default
risk, leverage( (DLTT+DLC) / AT) is included deﬁned as the ratio of long-term debt
plus short-term debt to the total assets. Furthermore, proﬁtability is added as a

control variable to take into consideration the ﬁrm’s ability to pay back the loan.
_ a:
e”dTN(’"(§’+§r-d+T’T) where d is the continuously compounded expected dividend yield,
and a is the expected volatility of the stock return. 7' is the continuously compounded risk-free
rate. P is the stock price at the end of the ﬁscal year. X is the strike price of the option, and T is
the maturity of the option.(see Edmans, Gabaix, and Landier (2008)) for more details)

lfiHljt = 21:, S.,-t where Sijt is the market share of ﬁrm i in industry j in year t. Firm’s
market share is computed using item #12 101

 

Proﬁtability is proxied with the ratio of EBITDA over Sales (OIBDP / Sales). All the
ﬁrm characteristic variables are one year prior to the year of loan origination.

Along with ﬁrm characteristics, I include loan—speciﬁc variables such as loan size,
maturity, and binary indicator variable, Secured, as in the previous empirical studies.
Loan size in its logarithm is included to control the size effect of the syndication.
Longer maturity also increases the default probability of a loan to the lenders, and
results in discouraging their participation. I include the logarithm of the loan maturity
(in terms of days). The presence of the Secured term in a contract can encourage more
lenders because when the loan becomes default, the creditors can, in some extent, limit
the loss of the loan. I also control for industry effects by including industry dummies
based on the two—digit SIC classiﬁcation that appears in the Dealscan data set, and
year ﬁxed effects by using dummies for the years of loan origination. The paper also
controls the loan purposes by using dummy variables for ﬁve categories. In all the
regressions, I report robust t-statistics that adjust for the clustering at the ﬁrm level.
To control the reputation effects of the lead arrangers, I construct the top 10 league
tables based on the original data sample for every year, and assign a binary value to
the loans when their lead arrangers appear in the league tables in the year of the loan

origination.

2.5 Results

In this section, I will show the results that are predicted in the previous sections and

discuss them based on the results.

2.5. 1 Primary results

I study the impact of the protection of creditor’s rights on the loan structure by esti-

mating the regressions with the Herﬁndahl and the shares of the loan held by the lead
102

 

arrangers as the dependent variables respectively. I control for ﬁrm’s characteristics,
loan features, year ﬁxed effects, and [industry ﬁxed effects to draw more robust infer-
ences between the dependent and the degree of creditor rights protection. Column (1)
and (2) in Table 2.2 show that the Herﬁndahl and the 70 held by lead are negatively
related to the G index, and the coefﬁcients are statistically signiﬁcant at 1%, and 5%
level respectively.

It presents that on average, when the ﬁrm employs stricter protection for its cred-
itors’ rights, the participant lenders are more willing to join in the loan syndicate,
resulting in the smaller shares held by the lead arrangers. To measure the economical
signiﬁcance of the results, I estimate the amount of decline in the Herﬁndahl when
the degree of protection measure is lowered by its one standard deviation. Declin-
ing by one standard deviation of the G index results in 5% lower in the Herﬁndahl
[(36.9*2.7)/1966.67]1.4

In columns (3) and (4), using two extreme portfolios: the ‘Dictatorship Portfolio’
of the ﬁrms with the strongest creditors’ rights(the G index 2 14), and the ‘Democ-
racy Portfolio’ of the ﬁrms with the weakest controls’ rights (the G index 3 5) (see,
Gompers, Ishii, and Metrick (2003)) as the independent variables, I estimate the co-
efﬁcient, B. Column (3) shows that the Dictatorship type ﬁrms experience decline
in the loan concentration by 224 when the dependent variable is the Herﬁndahl and
which amounts to a decrease by 11% on its average.

In columns (5) to (8), I break the G index down into smaller components. Columns
(5) and (6) report the results using the E index, and Column (7) and (8) the remaining
antitakeover provisions(computed as the difference between the G index and E index).
I ﬁnd that the E index is statistically insigniﬁcant but the difference between the G
index and E index is signiﬁcant at the 1% signiﬁcance level.

The loan concentrations are smaller when the ﬁrm size is larger. This is understood

 

lff the average ﬁrm takes stricter measure and become a Dictatorship type, the Herfindahl de-

crease by the two fold,[(36.9*2.7*2)/1966.67].
103

in the two different prolongs: information effect and default risk. Usually large ﬁrms
are more likely to ﬁnance externally when they expand their business and through
frequent outside ﬁnancing, the information on the ﬁrms are more publicly known to
the market. Such more and more transparent information may contribute to narrow-
ing information gap between the lead arrangers and the participant lenders. Secondly,
the large ﬁrms are less likely to default on their debts. Because of their large scope
of business, the ﬁrms may afford to defend themselves from the unexpected exter-
nal shock. Firm leverage is shown negatively related with the loan concentrations.
That the ﬁrms whose leverage is high implies that the ﬁrms are more at a risk of
bankruptcy. The lead arrangers try to co-insure the risk of bankruptcy by encourag-
ing more participation of lenders to avoid the case in which it is exposed to excess
risk if they take most of the loan. EBITDA/ Sales has a negative relationship with
the loan concentrations. That the ﬁrms are more proﬁtable, indicated by higher in
the EBITDA/ Sales, signals higher probability of loan repayment, resulting in lower
loan concentrations.

Loan characteristics are also shown to be signiﬁcant in explaining the loan con-
centration. As loan size increases, the loan concentration decreases. A large size
loan requires relatively more lenders to share the loan. Loan maturity has a negative
relationship with the loan concentration. As the loan maturity gets longer, the de-
fault risk of the loan also increases. For this reason, the lead arrangers attract more
lenders into the loan syndication to decrease their shares of the loan. The loan, which
is ‘Secured’, is shown to be more concentrated. The loan being secured implies two
opposite.interpretations: on one hand, that loan is secured implies less risk of losing
theloan for the lenders, thus it may result in more lenders’ participation. On the
other hand, that loan is secured may imply that the lead arrangers are required to
engage themselves in monitoring the collateral. This can signal more agency problems

between the lead arrangers and participant lenders. When the latter effect is more

104

apparent, then the fact that a loan is secured will rather discourage lenders from
participating, resulting in more loan concentration. The regression results attest that
the latter effect overrides the former effect.

I discuss that the ﬁrm’s protection of creditors’ rights affect the structure of syndi-
cated loans. However this impact might arise from the ease of takeover concern’ﬁavhen
the ﬁrms reinforce their antitakeover provisions (antitakeover channel). The following
empirical results thus are interpreted in the view of this channel. On the contrary,
this effect stems from the protection of creditor’s rights as I claim (protection chan-
nel). The task to distinguish which channel is more inﬂuential in terms of framing
the structure of loan syndicates is difﬁcult. However in the following subsections, I

want to provide some supporting evidence for the protection channel.

2.5.2 Risk-shifting and Loan structure

The risk-shifting problem may be more glaring in the situation when ﬁrms are ﬁ-
nancially distressedl.6 The shareholders are more tempted to make risk-increasing
investment when the ﬁrms are ﬁnancially distressed’if they can have the beneﬁts
when the investment is successful, whereas the creditors bear the costs when it turns
out poorly. In the time of lending decision, the creditors will be conscious of the pos-
sibility of such case and take some measures to guard themselves, especially when the
ﬁrms in question are ﬁnancially distressed. In the loan syndicates, it will be shown by
the lead arrangers’ holding more shares of a loan to be incentivised to step up their
monitoring efforts, or by including more covenants in the loan contracts.

The creditors can install various kinds of covenants in a loan. The number of
covenants thus indicate the severity of monitoring efforts by the creditors, especially

by the lead arrangers. This reasoning also implies that for the ﬁrms that provide

 

1Takeovers may entail increased ﬁnancial leverage, and debt seniority can be compromised
1Recent empirical study, Eisdorfer (2008) found risk-shifting behavior in the ﬁnancially distressed
ﬁrms

105

stronger protection of creditors’ rights, there will be less demand on the inclusion of
covenants.

Some syndicated loans also embed covenants which allow the lenders to intervene in
the management of the borrowing ﬁrm when the feasibility of borrower’s repayment is
doubted. if the covenants are not honored, the lenders can liquidate the project(the
ﬁrm) and claim the remaining of its assets, or enter into a renegotiation to renew
existing debts. Covenants may be used for err post mechanism of transferring control
rights from a borrowing ﬁrm to its creditors (Aghion and Bolton (1992), Hart and
Moore (1994)). When the lead arrangers include covenants in the loans, which will
send a'strong signal that there will be less severity of the adverse selection problem
and the lead arrangers are also committed to paying due diligence on the borrowers
(Boot, Greenbaum, and Thakor, 1993, Rajan and Winton, 1995)

To test whether the protection channel is more supported, I divide the sample
into two separate groups based on the Altman’s zscore(Altman, 1968). When the
ﬁrm’s zscore is below 1.81, then the ﬁrm is categorized into the distressed group,
and otherwise into the healthy group. The testing model is the same as in equation
(1.1) when the dependent variables are the Herﬁndahl and % Held by Lead. In
addition, I test the effect of the protection of creditors’ rights on the probability
of having collateral using probit model with a dummy dependent variable, Secured
while controlling variables are the same as before. Finally, Total number of ﬁnancial
couenantslﬂs taken as the dependent variable because of the importance of ﬁnancial
covenants in the loan contracts.

Table 2.3 reports the results of estimates of the G index coefﬁcient. In columns

(1) and (3), the G index is shown to affect the structure of syndicated loans twice as

 

1Unlike other covenants such as sweeps, ﬁnancial covenants provide simpler and more transparent
monitoring mechanism in that the monitoring on the ﬁrms by the lead arrangers, and the monitoring
on whether the lead arrangers pay due monitoring efforts by the participant lenders are relatively
easier because the ﬁrms’ ﬁnancial data are public and ﬁrms’ performance is veriﬁable with limited
costs.

106

much as the size of effect of the G index when the regression is estimated over the
whole sample and they are signiﬁcant at the 1% signiﬁcance level at both regressions.
However as seen in colurrms (2) and (4), the G index is not signiﬁcant for the sample
made of only healthy ﬁrms. These results are somewhat reinforced in the following
regressions. In columns (5) and (6), the G index is shown to affect the probability of
the creditors’ holding collateral at the 1% and 5% signiﬁcance level respectively. The
coefﬁcient of the G index in column (5) is about twice larger than that in column
(6). Furthermore, the results in columns (7) and (8) are also supportive in that the G
index affects how many ﬁnancial covenants the creditors embed in the loan contracts.
The stronger the ﬁrms’ protection measure of creditors’ rights, the fewer number of

ﬁnancial covenants the creditors include.

2.5.3 Asset Substitutions and Loan structure

In this section, I present more evidence in supporting the protection channel. The
creditors are aware of the possibility that the ﬁrms may substitute old assets of which
information the creditors(especially the lead arrangers in the loan syndicates) have
accumulated for new ones by investment when the ﬁrms have access to external ﬁ-
nancing opportunitiesl.8 This concern that the given creditors have may let them limit
the ﬁrms’ investment and growth by imposing stricter covenants. I also conjecture
that this asset substitution concern may affect the structure of syndicated loans, espe-
cially for the ﬁrms that are just able to have access to external ﬁnancing. Firms that
are not likely to have access to external ﬁnancing in the foreseeable future will have
less probability of the assets being replaced. However ﬁrms whose credit ratings are
just below investment grade may have a chance to receive external ﬁnancing. Thus

the concern about the asset substitution problem will be more felt for such ﬁrms. The

 

1fi’aulkender and Petersen (2006) show that when the ﬁrms have access to public ﬁnancial markets,
they take more debt

107

concern will be reﬂected also in the structure of syndicated loans. The participant
lenders, keenly aware of the possibility of the asset substitution, will be more likely
to prefer the ﬁrms whose measure of creditor rights protection is stronger, to the ones
that are not as much when the ﬁrms in question are the ones whose external ﬁnancing
window may just open.

To test this conjecture, I identify the relative likelihood of assets being substituted
for new ones by using ﬁrms’ credit ratings. Credit ratings provide information on
whether the ﬁrms have better access to external ﬁnancing opportunities. Among other
ratings, especially BB grade (in terms of S&P’s scale) is wedged between speculative
and investment demarcation. The ﬁrms whose rating is BB have more chances to
ﬁnance externally and by doing so substitute new assets for old ones compared to
the lower rated ﬁrms such as B, 000. Since the ﬁrms’ ratings are available before
loan origination, the participant lenders who concern about the asset substitution
may prefer the ﬁrms whose measure of creditor rights protection is stronger given the
same credit grades.

A testing regression model is extended from the equation (2.1),

5
Loan_Con,-jt = a + ”it + Z BkPCRJ-t x Ratingsjt + th6 + Xité + Eijt (2.2)
k=1

where PCR is a shortened notation for Protection of Creditors’ Right, which in-
teracts with the dummies (Ratings) representing each grade in AA, BBB, BB, B,
and 006’ 19respectively. The dummies take the value of one when the ﬁrms’ grade
belongs to either one of them respectively. In column (1) of Table 2.4, the coefﬁcient
of G index * BB is not only signiﬁcant at the 5% level, its size is the biggest across
the interaction terms. This result is also found in column (2) where the dependent

variable is the loan share held by the lead arrangers. These two results show that

 

1? drop the ﬁrms whose grades are either of the followings, AAA, CC, C, and D because of their
very small sample size. Out of 1954 rating matched loans, AA, A, BBB, BB, B, CCC grade ones
are 70, 470, 811, 458, 133, and 12 respectively

108

the participant lenders are more likely to lend to the ﬁrms whose measure of creditor

rights protection are stronger when there is a strong possibility of asset substitution.

2.5.4 CEO’s Compensation Structure

The alignment of manager’s compensation with its shareholders’ rights opens a possi-
bility of conﬂict between shareholders and creditors. Although the ﬁrms have a strong
protection measure for creditors’ rights, if the manager’s compensation structure is
aligned with the interest of the shareholders, the effectiveness of the protection will
be attenuated by the extent of the strength of the relationship. In addition, the G in-
dex is a composite index for measuring ﬁrm’s provisions with regard to antitakeover.
Thus the larger G index does not necessarily square with stronger protection of cred-
itors’ rights. Larger G index may imply stronger protection for the manager from the
replacement pressure of the shareholders. This behooves us to control the manager’s
incentive in the equation (1.1).

For the measures of the degree of the alignment, I use delta and vega 2.0 As seen
in columns ( 1) to (4) of Table 2.5, the G index still holds its signiﬁcance in affecting
the loan structure when the alignment is controlled either by the delta, or vega. The
coefﬁcients of the delta and vega are shown to be positive, which implies that as
the alignment between the manager and the shareholders becomes stronger, the lead
arrangers are asked to step up their monitoring efforts. The interpretation is carried
over when the dependent variable is Secured. However the decision on how many
ﬁnancial covenants the creditors embed in the loan contracts is affected neither by

the G index, nor the delta and vega.

 

2fnstcad, delta and vega, I use CEO’s stock ownership(%). However, it did not show any signiﬁ-
cant effect on the loan concentration

109

2.5.5 Impact of State’s Laws on Loan Contracts

I also study the impact of the state’s antitakeover measures on the loan concentra-
tion. To see how much the legal and institutional arrangements may contribute in
ameliorating the creditor rights infringement concern, ﬁrst I estimate the regressions
that have the severity of state’s protection for creditors’ rights, as the independent
variables, using Bebchuk and Cohen (2003)’s (henceforth BC) data. The states are
grouped into three categories by the degree of restrictiveness on hostile takeovers:
None-restrictive(BC measure = 0), Moderate( 1 S BC measure 5 3), and Restrictive
(BC measure 2 4).

The results in column (2) of Table 2.6 present a signiﬁcant negative relationship
between the loan share held by the lead arranger and state antitakeover measure at
the 5% signiﬁcance level. Compared with the loan whose borrower incorporates in
the state which does not take any measures of legal protection from hostile takeovers,
the ﬁrms incoporated in the states whose antitakeover measures are either moderate,
or restrictive, have less concentrated loans.

In addition to the study on the effects of state’s antitakeover provisions on the loan
concentration, the state’s payout restriction also is found to have an effect on the
requirement of collateral. The provision of collateral gives the lenders the opportunity
to seize the collateral, and by doing so, they may limit downside risks for the loans.
From the borrower’s point of view, collateral provision is committing itself to the
success of the project and can work for mitigating lender’s moral hazard concern on
the ﬁrm, such as liquidating some portion of the project for the beneﬁts of either
shareholders or managers: creditor rights may be compromised when the ﬁrm cashes
out liquid assets or repurchases its shares that may beneﬁt the shareholders. This
possible infringement on creditors’ rights may affect lenders’ contracting behavior to
deter the possibility. This may be expressed as requiring the borrower to provide

collateral against the loan. However if legal and institutional arrangements curbs
110

ﬁrm’s infringement of creditors’ rights for the beneﬁt of creditors, then there may be
less requirement for the inclusion of collateral in the contract.

One of the legal provisions in the state level in ensuring creditors’ rights is payout
restriction. The ﬁrm is required to satisfy the prerequisite requirement before it ex-
ecutes liquidation of assets and repurchase of its shares. In the state whose payout
restriction to which the ﬁrm is subject to, the ﬁrm is less pressured to provide collat-
eral to assuage the moral hazard concern raised by the lenders compared to the ﬁrm
whose incorporated state does not provide the law. The results in columns (3) and (4)
of Table 2.6 present that state’s strictest restriction(ratio of book capital / debt =1.25)
on payout contributes to decline in the incidence of collateral in a signiﬁcant level, by
45% in probability. Moderate level of restriction(the ratio =1) also make signiﬁcant
difference with the case of no restriction. In columns (5) and (6), to compare whether
the presence of restriction affect the provision of collateral, dummy variable, Payout
restriction(if the ratio is equal or greater than 1) is constructed and its coefficient
is estimated and found signiﬁcant at the 1% level. In columns (4) and (6), I also
controlled the case where the creditors included payout restriction covenant in the
loan contracts. Dealscan provides the binary information on the dividend covenant.

Overall, the results are very supportive that the lead arrangers are very much
beneﬁted from the state’s legal arrangements for protecting creditors’ rights. By
arranging the loans to the borrower whose incorporated state provides such legal
protection, the lead arrangers do not have to hold collateral and pay less monitoring

costs.

2.5.6 Product Market Competition

An intense product market competition provides corporate managers with incentives
to run ﬁrms with high efﬁciency because the disciplinary forces of competition remove

incompetent managers from the market. when the competition is high, the creditors
111

 

are more likely to face the possibility of conﬂicts with shareholders 2.1 To ﬁnd the
impact of the protection of creditors’ rights under the different degree of the product
market competition, I group the market competition into three tercile categories. I
estimate the regressions and have the results in columns (1) and (2) in Table 2.7,
the coefﬁcients of B are negative and signiﬁcant at the 1% and 5% level when the
competition level is high and meidum. when the dependent variable is the Herﬁndahl
index. These results are the same except when the competition level is high when the
dependent variable is % Held by Lead. The degree of the impact of the protection of
creditors’ rights declines as the competition becomes less severe. For the Dictatorship
type ﬁrms, when the competition is highest, the protection of creditors’ rights is most
pronounced. As the product market becomes less competitive, the protection of cred-
itors’ rights does not affect the loan share distribution among the lenders. Democracy
type ﬁrms do not show much difference across the competition environment.

The results show that the protection of creditors’ rights is effective when the prod-
uct market is competitive. The creditors are very aware of the risks of their rights
being compromised when the competition is high so that their decision of joining a
syndication is much affected whether the ﬁrm employs protective measures against
the risks. Collateral and ﬁnancial covenants22are another indicator how much the
lead arranger will pay monitoring efforts on the borrowers. Like the shares of the
loan held by the lead arrangers, when creditors’ rights are somewhat secured, they
are less likely to hold collateral or ﬁnancial covenants. This reason is attested by the
results in the regressions in columns (1), and (5) of Table 2.8. In column (1), when
the G index increases by one unit, the marginal effect rises by 2% and this is 5%

increase on the average.

 

2Grullon and Michaely (2005) observe that ﬁrms in less competitive industries have signiﬁcantly
lower payout ratios than ﬁrms in more competitive markets

2Financial covenants include the requirements that the borrower should maintain some ﬁnancial
condition measures at either minimum level, or maximum level: a minimum net worth, minimum
current ratio, minimum interest coverage, or not exceed a maximum leverage ratio.

112

 

When the G index is interacted with the market competition, the signiﬁcance of
the impact of creditor rights protection does not change across the competition levels
when the dependent variable is Secured. However, for the dependent variable, Number
of Financial Covenants, the impact is signiﬁcant when the market competition is high
and medium. Especially, for the Dictatorship type ﬁrms, the drop in the likelihood
of having collateral is pronounced when the ﬁrms belong to the high competition
industries. However, When the ﬁrms are the Democracy type, on the contrary, the

lead arrangers rather are more likely to hold collateral for securing against the risks.

2.6 Endogeneity

A concern that can arise when I use the G and E indices and the state legal provisions
is the endogeneity of those variables. The variables are also one of the factors that
the ﬁrms may consider when they think of loan ﬁnancing. The ﬁrms may tune the
level of monitoring they can incur while they also decide the level of creditor rights
protection (e.g., in which state it will incorporate). However, because the G index
shows consistency over the years, and its initial level is usually adopted at the time of
ﬁrms’ IPO (Chava, Livdan, and Purnanandam, 2009), the incidence of endogeneity

problem will be very limited.

2.7 Conclusion

The moral hazard problem of ﬁrms in syndicated loans declines when the participant
lenders are sure of the protection of their rights in the instances of interest conﬂicts
with the shareholders. I show that when the ﬁrms employ more measures for pro-
tecting creditors’ rights, the lead arrangers take a smaller portion of the loans, and in
consequence have less exposure to various risks. In addition, I found that when the

protection of creditors’ rights is institutiqriglized, the moral hazard problem between

the lead arrangers and the participant lenders decreases. Furthermore the legal and
institutional environments are shown to affect the probability of creditor’s holding
collateral and number of ﬁnancial covenants. The effectiveness of the protection of
creditor’s rights also depends on the product market competition. The ﬁrms which
belong to competitive industries have the less concentrated loans when their pro—
tection of creditors’ rights are very strong and also less likely provide collateral and

ﬁnancial covenants.

114

Figure 2 Relationship with the G index and the loan concentration measures

Lowess smoother

 

 

 

 

am

8
22 cases

mmm5><

bandwidth = .3

Lowess smoother

 

 

 

0000 _.

88 88 DDS Sam
.32.. cmeceisﬁucesr

=.8

bandwidth

115

 

 

 

 

 

Table 2.1 Descriptive Statistics
Panel A
G index <=5 6 7 8 9 10 11 12 13 >=14 All

Number of Loans 347 284 356 406 473 475 386 247 180 162 3316
Panel B
Variable Mean STD. 25th 50th 75th 95th
Loan characteristics

Loan size (5 Million 518 1074 100 250 500 1000
Maturity(days) 1325 728 738 1213 1826 1840
Secured [0,1] 0.44 0.50 0 0 1 1
Tm“ ’lum’m °’ 2.50 1 .10 2 2 3 4
ﬁnanc1al covenants

Net worth covenant 0.43 0.49 0 0 1 1
Loan structure

% Held by Lead 22.60 17.00 10.19 17.32 30.00 50.00
Hfmndah’ 1948.6 1671.3 795.0 1357.5 2512.0 4961.8
-H1rschman
Firm characbristics

G index 9.03 2.72 7 9 1 1 13
E index 2.35 1.32 1 2 3 4
Total_assets
(S Billions) 4.60 10.71 0.46 1.29 3.73 12.23
Leverage 0.30 0.21 0.16 0.28 0.40 0.52
EBITDA/Sales 0.17 0.16 0.09 0.14 0.22 0.33
Zscore 1.82 1.22 1.01 1.77 2.57 3.85
Governance

CEO equity share(% 2.12 5.50 0.00 0.00 1.20 6.53
Delta (S in '000) 1057.1 5699.0 67.1 192.7 5521 1762.7
Vega (S in '000) 141.7 330.5 6.9 38.2 125.6 345.0

 

116

Table 2.2 Creditor's Rights and Loan Structure

This table reports results from the loan structure regressions that use corporate charter
provisions and state-level antitakeover restrictions as proxies for creditor's rights. The dependent
variables are the average share held by the lead arranger(sXmax < 100), and a constructed loan
concentration using lenders' loan share, Herﬁndahl-Hirschman index(max <10,000). G index
from Gompers et al. (2003) is used as a primary proxy for creditor's rights. Firms are grouped
based on the G index into Dictatorship when its G index is larger than 13 and Democracy when
it is below 6. "G index minus E index" is the remaining provisions in the G index after
excluding the E index in Bebchuk et al.(2009). The explanatory variables include ﬁnn-speciﬁc
and loan-speciﬁc variables. ln(total assets) refers to the logarithm of total assets of the ﬁrm in
millions of USD before the loan initialization. ebitda/total assets is the ratio of EBITDA of the
ﬁrm to the total assets of the ﬁrm. leverage is the total debt (long term plus short term) divided
by the total assets of the ﬁrm. log(deal amount) refers to the logarithm of the size of the loan.
ln(matun'ty) is the logarithm of the days of loan maturity. "Secured" is the binary variable,
which is a unit of one when the loan is secured and lead arranger’s reputation is controlled by the
binary variable, 'Top 10 leader" having a unit of one when the lead arranger is top 10 ranked in
the prior year to the year of loan origination. Standard errors are estimated with clustered errors
at the ﬁrm level and are reported in parentheses. Regressions also include industry ﬁxed
effects(2 digit SIC), year ﬁxed effects, and loan purpose indicators. *, *"‘ indicate the level of
signiﬁcance at 5% and 1%, respectively.

 

 

Herﬁndahl % Held by Lead Herﬁndahl % Held by Lead
(1) (2) (3) (4)
G index -36.92""" -0.29*
(11.85) (0.13)
Dictatorship -224.30* -1 .91
(96.07) (1.30)
Democracy 65.28 1 .05
(119.32) (1.28)
Ln(total assets) 13.24 -1.44** 3.36 -1 .50“I
(47.26) (0.49) (47. 13) (0.48)
Leverage -317.98* -4.93"‘ -314.32 -4.95"‘
(159.75) (1.91) (162.67) (1.94)
Ebitda/Sales -670.13"”" -3.53 -663.85"‘* -3.45
(209.42) (2.36) (213.79) (2.39)
Ln(deal size) -728.42“ —6.36" -724.92"“" -6.34**
(58.82) (0.60) (58.98) (0.60)
Ln(maturity) -238.32" -3.39"‘* -237.42** -3.38*"
(61.48) (0.61) (61.81) (0.61)
Secured 230.73" 2.00“ 251.41" 2.14"
(73.97) (0.79) (75.23) (0.80)
Top 10 lender -l64.25"' -l.56"I -160.56* -l.53"‘
(67.00) (0.66) (67.22) (0.66)
Intercept 17,816.72" 179.99" 17,457.41" 177.25"
(909.33) (9.46) (925.87) (9.52)
N 2155 2155 2155 2155
R-sqwed 0.40 0.41 0.40 0.40

 

117

. Table 2.2 Creditor's Rights and Loan Structure (Cont.)

This table reports results from the loan structure regressions that use corporate charter
provisions and state-level antitakeover restrictions as proxies for creditor's rights. The dependent
variables are the average share held by the lead arranger(sXmax < 100), and a constructed loan
concentration using lenders' loan share, Herﬁndahl-Hirschman index(max <10,000). G index
from Gompers et a1. (2003) is used as a primary proxy for creditor's rights. Firms are grouped
based on the G index into Dictatorship when its G index is larger than 13 and Democracy when
it is below 6. "G index minus E index" is the remaining provisions in the G index after
excluding the E index in Bebchuk et al.(2009). The explanatory variables include ﬁrm-speciﬁc
and loan-speciﬁc variables. ln(total assets) refers to the logarithm of total assets of the ﬁrm in
millions of USD before the loan initialization. ebitda/total assets is the ratio of EBITDA of the
ﬁrm to the total assets of the ﬁrm. leverage is the total debt (long term plus short term) divided
by the total assets of the ﬁrm. log(deal amount) refers to the logarithm of the size of the loan.
ln(maturity) is the logarithm of the days of loan maturity. "Secured" is the binary variable,
which is a unit of one when the loan is secured and lead arranger's reputation is controlled by the
binary variable, 'Top 10 leader" having a unit of one when the lead arranger is top 10 ranked in
the prior year to the year of loan origination. Standard errors are estimated with clustered errors
at the ﬁrm level and are reported in parentheses. Regressions also include industry ﬁxed
effects(2 digit SIC), year ﬁxed effects, and loan purpose indicators. ’, "“' indicate the level of
gigniﬁcance at 5% and 1%, respectively.

 

 

Herﬁndahl % Held by Lead Herﬁndahl % Held by Lead
(5) (6) (7) (8)
E index -38.53 -0.41
(24.73) (0.25)
G index-E index -46.88"”" -0.33
(15.37) (0.18)
Ln(total assets) 3.41 -1.76"‘* 14.33 4.44"
(45.79) (0.48) (47.37) (0.49)
Leverage -344.83"' -5.17** -309.41 -4.87*
(155.71) (1.85) (160.80) (1.93)
Ebitda/Sales -640.43‘“‘l -3.35 -678.17"”" -3.58
(216.24) (2.37) (210.42) (2.37)
Ln(deal size) 435.06” -6.24"”" -725.24"“" -6.33"”"
(57.10) (0.59) (58.90) (0.60)
Ln(maturity) -243.4l ** -3.52** ~239. 16" -3.40"
(60.86) (0.59) (61.41) (0.61 )
Secured 236.62" 184* 236.67" 2.06"
(73.55) (0.78) (74.20) (0.79)
Top 10 lender -158.34"‘ -1.37* 466.13" -1.57"‘
(65.60) (0.65) (67.09) (0.66)
Intercept 17,786.93" 179.43" 17,729.76" 179.07"
(892.27) (9.30) (916.65) (9.50)
N 2227 2227 2155 2155
R-sguared 0.40 0.41 0.40 0.40

 

118

Table 2.3 Risk-Shifting and Loan Structure

This table reports results from the loan structure regressions that use corporate charter provisions(G
index from Gompers et al. (2003)) as a proxy for creditor's rights. The dependent variable in column
( 1) and (2) is the average share held by the lead arranger(s)(max < 100), in column (3) and (4), a
constructed loan concentration using lenders' loan share, Herﬁndahl-Hirschman index(max <10,000)
In column (5) and (6), the dependent variable is a binary one, "Secured", which takes the value of
one when the loan is secured with collateral, zero otherwise. In column (7) and (8), "Number of
ﬁnancial covenans" is the dependent variable which is total number of ﬁnancial covenants that the
loan includes Firms are grouped into Distressed ﬁrms when Altman's zscore is below 1.81, and into
Healthy ﬁrms when the score is greater than, or equal to 1.81. The explanatory variables include
ﬁrm-speciﬁc and loan-speciﬁc variables. ln(total assets) refers to the logarithm of the total assets of
the ﬁrm in millions of USD before the loan initialization. ebitda/total assets is the ratio of EBITDA
of the ﬁrm to the total assets of the ﬁrm. leverage is the total debt (long term plus short term) divided
by the total assets of the ﬁrm. log(dea1 amount) refers to the logarithm of the size of the loan.
ln(maturity) is the logarithm of the days of loan maturity. "Top 10 leader" having a unit of one when
the lead arranger is top 10 ranked in the prior year to the year of loan origination. Standard errors are
estimated with clustered errors at the ﬁrm level and are reported in parentheses. Regressions also
include industry ﬁxed effects(2 digit SIC), year ﬁxed effects, and loan purpose indicators. *, "
indicate the level of signiﬁcance at5% and 1%, respectively.

 

 

 

ﬁerﬁndam % Held by Lead
Distressed ﬁrms Healthy ﬁrms Distressed ﬁrms Healthy ﬁrms
(1) (2) £3) (4)
G index -72.80*"' -0.47 -0.59" —0.07
(18.49) (15.28) (0.20) (0.18)
Ln(total assets) 20.71 41.09 -1.14 -1.26
(66.10) (65 .66) (0.69) (0.72)
Leverage -98.69 -538.50" -1.29 -9.03"""
(245.56) (236.25) (2.62) (2.91)
Ebitda/Sales -701.98""" -518.57 -5.30"' 0.16
(240.07) (473.60) (2.64) (5. 12)
Ln(deal size) -636.97" -874.08*"' -6.29*"' -7.08"
(70.23) (90.51) (0.74) (0.98)
Ln(maturity) -142.28 -285.24"‘"' -2.90" -3.72"
(92.61) (81.77) (0.88) (0.87)
Secured 326.05" 94.59 1.50 2.65“
(106.41) (95.02) (1.22) (1.03)
Top 10 lender -238.17* -100.65 -2.23"' -0.70
(1 10.15) (74.79) (1.03) (0.84)
Intercept 16,403.13" 20,153.36“I 179.62" 18828"
(1380.64) (1342.16) (12.92) (14.84)
N 1026 1129 1026 1129
R-sguared 0.38 0.51 0.41 0.45

 

119

Table 2.3 Risk-Shifting and Loan Structure (cont.)

This table reports results from the loan structure regressions that use corporate charter provisions(G
index from Gompers et al. (2003)) as a proxy for creditor's rights. The dependent variable in column
(1) and (2) is the average share held by the lead arranger(s)(max < 100), in column (3) and (4), a
constructed loan concentration using lenders' loan share, Herﬁndahl-Hirschman index(max <10,000).
In column (5) and (6), the dependent variable is a binary one, "Secured", which takes the value of
one when the loan is secured with collateral, zero otherwise. In column (7) and (8), "Number of
ﬁnancial covenants" is the dependent variable which is total number of ﬁnancial covenants that the
loan includes. Firms are grouped into Distressed ﬁrms when Altman's zscore is below 1.81 , and into
Healthy ﬁrms when the score is greater than, or equal to 1.81. The explanabry variables include
ﬁrm-speciﬁc and loan-speciﬁc variables. ln(total assets) refers to the logarithm of the total assets of
the ﬁrm in millions of USD before the loan initialization. ebitda/total assets is the ratio of EBITDA
of the ﬁrm to the total assets of the ﬁrm. leverage is the total debt (long term plus short term) divided
by the total assets of the ﬁrm. log(dea1 amount) refers to the logarithm of the size of the loan.
ln(maturity) is the logarithm of the days of loan maturity. "Top 10 leader" having a unit of one when
the lead arranger is top 10 ranked in the prior year to the year of loan origination. Standard errors are
estimated with clustered errors at the ﬁrm level and are reported in parentheses. Regressions also
include industry ﬁxed effects(2 digit SIC), year ﬁxed effects, and loan purpose indicators. ‘, *"
indicate the level of si niﬁcance at 5% and 1%, re ectivel .

Number of Financial Covenants

 

 

Distressed ﬁrms Healthy ﬁrms Distressed ﬁrms Healthy ﬁrms
(5) (6) (7) (8)
G index -0.1 1" -0.05“' -0.07""" 0.01
(0.02) (0.02) (0.01) (0.01)
Ln(total assets) -0.20"'* -0.43*"‘ -0.1 1'" -0.23**
(0.06) (0.08) (0.04) (0.05)
Leverage 0.81 ” 1.63“ "' -0.08 0.20
(0.31) (0.33) (0.18) (0.20)
Ebitda/Sales -0.95"' -1.74" -O.25 -0.31
(0.44) (0.72) (0.15) (0.45)
Ln(deal size) -0.25** 0.00 -0.12" -0.10
(0.07) (0.09) (0.05) (0.06)
Ln(maturity) 0.07 0.02 0.20" 0.05
(0.09) (0.09) (0.06) (0.06)
Secured - - 0.47" 0.32"
- - (0.09) (0.08)
Top 10 lender -0.12 0.07 019" 0.00
(0.1 1) (0.1 1) (0.07) (0.06)
Intercept 10.81" 2.61 3.32" 6.02”
(1.39) (1.48) (0.91) (0.99)
N 993 1102 823 936
R-squared - - 0.52 0.42

 

120

Table 2.4 Asset Substitutions and Loan Structure

This table reports results from the loan structure regressions that use corporate charter
provisions(G index from Gompers et al. (2003)) as a proxy for creditor’s rights. The dependent
variable is the average share held by the lead arranger(s)(max < 100), a constructed loan
concentration using lenders' loan share, Herﬁndahl-Hirschman index(max <10,000), "Secured",
which takes the value of one when the loan is secured with collateral, zero otherwise, and
"Number of ﬁnancial covenants" which is total number of ﬁnancial covenants that the loan
includes. The explanatory variables include ﬁrm-speciﬁc and loan-speciﬁc variables. Firms'
credit ratings are denoted by AA, A, BBB, BB, B, and CCC. ln(total assets) refers to the
logarithm of the total assets of the ﬁrm in millions of USD before the loan initialization.
ebitda/total assets is the ratio of EBITDA of the ﬁrm to the total assets of the ﬁrm. leverage is
the total debt (long term plus short term) divided by the total assets of the ﬁrm. log(deal amount)
refers to the logarithm of the size of the loan. ln(maturity) is the logarithm of the days of loan .
maturity. "Top 10 leader" having a unit of one when the lead arranger is top 10 ranked in the
prior year to the year of loan origination. Standard errors are estimated with clustered errors at
the ﬁrm level and are reported in parentheses. Regressions also include industry ﬁxed effects(2
digit SIC), year ﬁxed effects, and loan purpose indicators. *, ** indicate the level of signiﬁcance
at 5% and 1%, respectively.

Herﬁndahl % Held by Lead Secured Number of Financial Covenants

 

 

G index “ AA -35.12 -0.15 -0.06 -0.07"”"
(18.82) (0.20) (0.04) (0.02)
G-index "' A -29.65 -0.14 -0.16** -0.06"“"
(15.15) (0.15) (0.03) (0.01)
G index * BBB -40.67"”" -0.23 -0.14"'* -0.01
(15.49) (0.16) (0.02) (0.01)
G index * BB -45.33"‘ -0.36* 0.03 ' 0.01
(18.06) (0.17) (0.02) (0.02)
G index * B 82.79“ 0.80’ 0.07 0.00
(34.41) (0.34) (0.04) (0.02)
G index * CCC 79.53 -0.25 - 0.00
(65.99) (0.70) - (0.07)
Ln(total assets) 123.91 0.05 -O.13 -0.1 1*
(63.97) (0.60) (0.07) (0.04)
Leverage 55.81 0.52 0.65“ 0.03
(261.41) (2.98) (0.33) (0.19)
Ebitda/ Sales -2 1 3 .01 -2.54 -0.13 -0.06
(320.16) (3.37) (0.59) (0.36)
Ln(deal size) -559.53** -4.90"”" -0.18‘l -0.07
(74.77) (0.69) (0.07) (0.05)
Ln(maturity) -73.89 -2.22"”" 0. 16 0.09"
(85.39) (0.76) (0.09) (0.04)
Secured 221.12 2.56“ - 0.44"
(123.27) (1.25) - (0.09)
Top 10 lender -96.76 -1.00 0.15 -0.03
(83.04) (0.76) (0.1 1) (0.06)
Intercept 11,654.63" 125.62" 4.36" 2.61"
-l306.22 -12.33 -1.63 -0.89
N 1277 1277 1241 1068
R-squared 0.3 14 0.318 - 0.492

 

121

Table 2.5 CEO compensation and Loan Structure

This table reports results from the loan structure regressions that use corporate charter provisions
(G index from Gompers et al. (2003)) as a proxy for creditor’s rights. The dependent variable in
column (1) and (2) is the average share held by the lead arranger(sXmax < 100), in column (3) and
(4), a constructed loan concentration using lenders' loan share, Herﬁndahl-Hirschman index(max
<10,000). In column (5) and (6), the dependent variable is a binary one, "Secured", which takes the
value of one when the loan is secured with collateral, zero otherwise. In column (7) and (8),
"Number of ﬁnancial covenants" is the dependent variable which is total number of ﬁnancial
covenants that the loan includes. Delta is deﬁned as the dollar change in a CEO's stock and option
portfolio for a 1% change in stock price and Vega, as the dollar change in a CEO's option holdings
for a 1% change in stock return volatility. The explanatory variables include ﬁrm-speciﬁc and loan-
speciﬁc variables. ln(total assets) refers to the logarithm of total assets of the ﬁrm in millions of
USD before the loan initialization. ebitda/total assets is the ratio of EBITDA of the ﬁrm to the total
assets of the ﬁrm. leverage is the total debt (long term plus short term) divided by total assets of the
ﬁrm. log(deal amount) refers to the logarithm of the size of the loan. ln(maturity) is the logarithm
of the days of loan maturity. "Top 10 leader" having a unit of one when the lead arranger is top 10
ranked in the prior year to the year of loan origination. The coefﬁcients are estimated using probit
model and standard errors are reported in parentheses and clustered by ﬁrm. Regressions also
include industry ﬁxed effects(2 digit SIC), year ﬁxed effects, and loan purpose indicators. "', "”"
indicate the level of signiﬁcance at 5% and 1%, respectively.

 

 

 

Herﬁndahl %Held by Lead Secured . Nlm‘be’ °f

Frnancral Covenants

(1) (2) (3) (4) (5) (6) (7) (8)

Gindex -40.74"”'I -42.31""l -0.35* -0.36* -0.07" -0.07*"' -0.02 -0.02
(13.45) (13.39) (0.14) (0.14) (0.02) (0.02) (0.01) (0.01)

Delta 7.41 "' 0.06 -0.01 0.00
(3.15) (0.04) (0.02) (0.00)

Vega 274.59" 2.73""I -0.56"‘ -0.1 1
(83.16) (0.85) (0.28) (0.08)
Ln(total asset: -5.02 -23.46 -1.21* -1.40"”" -0.23""l -0.20" -0.16"”" -0.15"
' (48.30) (49. 18) (0.51) (0.52) (0.05) (0.06) (0.04) (0.04)

Leverage -74.54 -57.55 -2. 15 -1.94 2.01 " 1.94" 0.18 0.17
(205.12) (202.72) (2.65) (2.61) (0.31) (0.31) (0.16) (0.16)

Ebitda/Sales -556.90 -635.57* —5.07 -5.89 -1.62" -1.53"”" -0.22 -0.20
(284.77) (283.73) (3.08) (3.09) (0.46) (0.47) (0.27) (0.27)
Ln(deal size) -678.08** -682.47" -6.50" -6.56"”" -0.18** -0.18" -0.11” -0.11"
(57.43) (57. 12) (0.62) (0.62) (0.06) (0.06) (0.04) (0.04)

Ln(maturity) 499.97" 494.67" -2.16*"‘ -2.10** 0.11 0.11 0.12" 0.12"
(67.92) (67.48) (0.66) (0.66) (0.07) (0.07) (0.04) (0.04)

Secured 290.52" 298.38" 2.30" 2.38" 0.45" 0.44"
(77. 10) (77.02) (0.87) (0.87) (0.06) (0.06)

Top lOlender -139.45 -131.75 -1.45* -1.38 -0.02 -0.02 -0.06 -0.07
(74.81) (74.74) (0.74) (0.74) (0.08) (0.08) (0.05) (0.05)

Intercept 20,358.82**20,622.78"211.46*‘214.35" 5.56" 5.31" 4.49” 4.41"
(984.35) (997.00) (10.76) (10.86) (1.13) (1.14) (0.96) (0.96)

N 1704 1704 1704 1704 1688 1688 1447 1447

R-squgred 0.39 0.39 0.39 0.40 0.44 0.44

 

122

Table 2.6 State Law and Loan Structure
This table reports results from the loan structure regressions that use state-level antitakeover
provisions and payout restriction as proxies for creditor's rights. The dependent variables in
column (1) and (2) are the average share held by the lead arranger(s)(max < 100), and a
constructed loan concentration using lenders' loan share, Herﬁndahl-Hirschman index(max
<10,000). In column (3) to (6), the dependent variable is a binary one, "Secured", which takes
the value of one when the loan is secured with collateral, zero otherwise. State-leve antitakeover
index based on state's ﬁve antitakeover provisions (Bebchuck and Cohen, 2003) is categorized
into three groups ("Not restrictive" when the index is zero, "Moderate" when the index falls
between 1 and 3, and "Restrictive" larger than 3). "Not restrictive" is omitted in the regression as
a base group. Some states also impose restriction on ﬁnn's payout activity based on the ratio of
equity to debt. "Payout moderate" states ask the ﬁrms to have the ratio at minimun one. "Payout
restrictive" at minimum 1.25. "Payout restriction" is a binary variable having the value of one
when there is a state-level restriction on ﬁrm's payout The explanatory variables include ﬁrm-
speciﬁc and loan-speciﬁc variables. ln(total assets) refers to the logarithm of the total assets of
the ﬁrm in millions of USD before the loan initialization. ebitda/total assets is the ratio of
EBITDA of the ﬁrm to the total assets of the ﬁrm. leverage is the total debt (long term plus short
term) divided by the total assets of the ﬁrm. Iog(deal amount) refers to the logarithm of the size
of the loan. ln(maturity) is the logarithm of the days of loan maturity. "Top 10 leader" having a
unit of one when the lead arranger is top 10 ranked in the prior year to the year of loan
origination. The coefﬁcients are estimated using probit model and standard errors are reported in
parentheses and clustered by ﬁrm. Regressions also include industry ﬁxed effects(2 digit SIC),
year ﬁxed effects, and loan purpose indicators. *, " indicate the level of signiﬁcance at 5% and

 

 

 

Herﬁndahl % Held by Lead Secured
(l) (2) (3) (4) (5) (6)
Moderate State -164.37 -4.24"'
(209.58) (1.93)
Restrictive State -172.34 -4.15"'
(216.26) (1.99)
Payout moderate -0.32" -0.27"”"
(0.09) (0.09)
Payout restrictive -1.13" -1.26”
(0.32) (0.34)
Payout restriction -0.34** -0.30"”"
(0.09) ( 0.09)
Ln(total assets) 13.68 4.62" -0.32” -0.33" -0.32" -0.32"
(45.87) (0.49) (0.05) (0.05) (0.05) (0.05)
Leverage -349.23* -5.01** 1.46" L68" “ 1.46“ * 1.67"
(157.80) (1.87) . (0.24) (0.32) (0.24) (0.31)
Ebitda/Sales -645.91" -3.86 -l.l7"“" -0.99"‘ -1.18" -l.01"'
(218.47) (2.30) (0.36) (0.40) (0.36) (0.40)
Ln(deal size) -749.41** -6.47“'"‘ -0.14* "‘ -0. 13" -0.15"“" -0. 13"
(57.75) (0.59) (0.05) (0.06) (0.05) (0.06)
Ln(maturity) -229.43" -3.37** 0.04 0.08 0.04 0.07
(61.06) (0.59) (0.06) (0.07) (0.06) (0.07)
Secured 251 .45" * 2.03“ * - - - -
(73.17) (0.78) - - - -
Dividend Coven: - - - 0.76" - 0.75"
- - - (0.10) - (0.10)
Top 10 lender -16l.75"‘ o1.44* -0.04 -0.10 -0.04 ~0.11
(66.59) (0.66) (0.07) (0.08) (0.07) (0.08)
Intercept 17,942.33" 184.48" 4.63" 4.37" 4.67" 4.45"
(929.91) (9.52) (1.00) (1.23) (1.00) (1.23)
N 2192 2192 2179 1889 2179 1889
R-sguared 0.40 0.42 - - - -

 

123

Table 2.7 Market Competition and Loan Structure

This table reports results from the loan structure regressions that use corporate charter provisions (G
index from Gompers et al. (2003)) as a proxy for creditor's rights. The dependent variables are the
average share held by the lead arranger(s)(max < 100), and a constructed loan concentration using
lenders' loan share, Herﬁndahl-Hirschman index(max <10,000). Firms are grouped based on the G
index into Dictatorship when its G index is larger than 13 and Democracy when it is below 6. HHI is
also a Herﬁndahl-Hirschman index for product market competition. It is cateorized into three groups,
lowest, medium and high tercile of its empirical distribution. The explanatory variables include ﬁrm-
speciﬁc and loan-speciﬁc variables. ln(total assets) refers to the logarithm of the total assets of the ﬁrm
in millions of USD before the loan initialization. ebitda/total assets is the ratio of EBITDA of the ﬁrm
to the total assets of the ﬁrm. leverage is the total debt (long term plus short term) divided by total
assets of the ﬁrm. log(deal amount) refers to the logarithm of the size of the loan. ln(maturity) is the
logarithm of the days of loan maturity. "Secured" is the binary variable, which is a unit of one when the
loan is secured and lead arranger’s reputation is controlled by the binary variable, ‘Top 10 leader"
having a unit of one when the lead arranger is top 10 ranked in the prior year to the year of loan
origination. The coefﬁcients are estimated using probit model and standard errors are reported in
parentheses and clustered by ﬁrm. Regressions also include industry ﬁxed effects(2 digit SIC), year
ﬁxed effects, and loan purpose indicators. *, *" indicate the level of signiﬁcance at 5% and 1%,
respectively.

 

Herﬁndahl% Held by LeadHerﬁndahl% Held by LeadHerﬁndahl% Held by Lead

 

(1) (2) (3) (4) (5) (6)
G index‘HHl(low) -40.66" -0.38"
(13.42) (0.15)
G index*HHI(medium) 49.02" -0.36"'
(13.31) (0.15)
G index*HHI(high) -24.05 -0. 17
(13.65) (0.15)
Dictatorship*HHI(low) -289.82" -3.82"'
(137.39) (1.53)
Dictatorship'HHKmedium) -189.06 -0.20
(184.98) (2.85)
Dictatorship’HHI(high) -210.49 -1 .90
(136.97) (1.81)
Democracy‘HHIUow) -45.41 -1 .46
(259.67) (2.52)
Democracy‘HHI(medium) 14.62 0.97
(206.21) (2.16)
Democracy‘HHI(high) 210.74 2.89
(167.87) (1.72)
Ln(total assets) 8.57 -1.45"”" 0.47 -1 .54" 2.20 - l .49"
(47.42) (0.49) (47.32) (0.49) (46.93) (0.48)
Leverage -303.75 499" -305.04 -4.9 l * -303.19 -4.87“
(159.24) (1.92) (161.92) (1.94) (163.00) (1.96)
Ebitda/Sales -661.33** -3.39 -674.52" -3.48 -663 .44" -3 .46
(211.06) (2.39) (214.94) (2.43) (220.91) (2.52)

124

Table 2.7 Market Competition and Loan Structure (cont.)

Ln(deal size) -72239" -6.32** -722.62"”" -6.32** -72392" -6.35**
(59.05) (0.60) (59.1 1) (0.60) (58.94) (0.59)
Ln(maturity) -241.65** -339" 239.01" -337" 241.39" -3.38**
(61.49) (0.61) (62.01) (0.61) (61.90) (0.61)
Secured [0,1] 226.25” 199* 254.00" 2.17M 257.97M 2.22"
(73.94) (0.80) (75.39) (0.80) (76.00) (0.80)
Top 10 lender [0,1] -167.59* -1.58* 459.4411 -1.50* 461.92" -1.55*
(67.41) (0.66) (67.22) (0.66) (66.94) (0.66)
177.88 176.86 177.30
Intercept 17625.24 ,, 17437.16 ,, 17480.17 ,,
#1! It **
(914.45) (9.58) (929.25) (9.57) (926.15) (9.51)
N 2150 2150 2150 2150 2150 2150
R-squared 0.40 0.41 0.40 0.40 0.40 0.40

 

125

Table 2.8 Market Competition and Loan Structure (Cont.)

This table reports results from the loan structure regressions that use corporate charter provisions
(G index from Gompers et al. (2003)) as a proxy for creditor's rights. In column (1) to (4), the
dependent variable is a binary one, "Secured", which takes the value of one when the loan is
secured with collateral, zero otherwise. In column (5) to (8), "Number of ﬁnancial covenants" is
the dependent variable which is total number of ﬁnancial covenants that the loan includes. F inns
are grouped based on the G index into Dictatorship when its G index is larger than 13 and
Democracy when it is below 6. HHI is also Herﬁndahl-Hirschman index for product market
competition. It is cateorized into three groups, lowest, medium and high tercile of its empirical
distribution. The explanatory variables include ﬁrm-speciﬁc and loan-speciﬁc variables. ln(total
assets) refers to the logarithm of total assets of the ﬁrm in millions of USD before the loan
initialization. ebitda/total assets is the ratio of EBITDA of the ﬁrm to the total assets of the ﬁrm.
leverage is the total debt (long term plus short term) divided by total assets of the ﬁrm. log(deal
amount) refers to the logarithm of the size of the loan. ln(maturity) is the logarithm of the days of
loan maturity. "Top 10 leader" having a unit of one when the lead arranger is top 10 ranked in the
prior year to the year of loan origination. The coefﬁcients are estimated using probit model and
standard errors are reported in parentheses and clustered by ﬁrm. Regressions also include
industry ﬁxed effects(2 digit SIC), year ﬁxed effects, and loan purpose indicators. *, " indicate
the level of signiﬁcance at 5% and 1%, respectively.

 

 

 

Secured Number of Financial Covenants
(1) (2) (3) (4) (5) (6) (7) (8)
G index -0.07"'"‘ -0.02"‘
p (0.02) (0.10)
G index * HHI(low) -0.05** -0.04**
(0.02) (0.01)
G index * HHI(medium) -0.08** -0.02*
(0.02) (0.01)
G index "‘ HHI(high) -0.07"”" 0.00
(0.02) (0-01)
Dictatorship "' HHI(low) -0.66* -0.23*
(0.27) (0.1 1)
Dictatorship * HHI(medium) -0.24 -0.18
(0.27) (0.19)
Dictatorship "‘ HHI(high) 0.12 0.35
(0.24) (0.25)
Democracy " HHI(low) 0.73" ‘ 0.31"
(0-25) (0.14)
Democracy * HHI(medium) 0.28 0.24
(0.21) (0.14)
Democracy "' HHI(high) 0.11 ~0.15
(0.20) (0.13)

Ln(total assets) -0.26"”" -027" _030... -029" -017" -017" -0.18"”" -0.18**
(0.05) (0.05) (0.05) (0.05) (0.03) (0.03) (0.03) (0.03)
Leverage 1.40M 1.44" 1.41" 1.40" 0.15 0.13 0.15 0.15
(0.23) (0.23) (0.23) (0.23) (0.13) (0.13) (0.12) (0.12)

126

Table 2.8 Market Competition and Loan Structure (Cont.)

Ebitda/Sales -1.28** -1.28** -1.30** -1.26** 035* 033* 035* 033*
(0.36) (0.36) (0.36) (0.36) (0.15) (0.15) (0.15) (0.15)

Ln(deal size) 015** 014** 013* 014* 009* 009* 009* 009*
(0.05) (0.05) (0.06) (0.05) (0.04) (0.04) (0.04) (0.04)

Ln(maturity) 0.04 0.03 0.03 0.03 0.11** 0.12** 0.11** 0.11**
(0.06) (0.06) (0.06) (0.06) (0.04) (0.04) (0.04) (0.04)

Top 101ender [0,1] 002 002 001 001 008 009 008 -0.08
(0.07) (0.07) (0.07) (0.07) (0.05) (0.05) (0.05) (0.05)

Secured [0,1] - - - — 0.42** 0.43** 0.43** 0.43**
- - - - (0.06) (0.06) (0.06) (0.06)

Intercept 4.89** 4.95** 4.22** 4.27** 3.82** 4.20** 4.15** 4.28**
(1.01) (1.01) (1.03) (1.02) (0.88) (0.82) (0.81) (0.81)

N 2144 2138 2138 2138 1759 1755 1755 1755
R-squared - - - — 0.43 0.43 0.43 0.43

 

127

2.8 Bibliography

AGHION, B, AND P. BOLTON (1992): “An incomplete contracts approach to ﬁnan-
cial contracting,” The Review of Economic Studies, 59(3), 473—494.

ALTMAN, E. (1968): “Financial ratios, discriminant analysis and the prediction of
corporate bankruptcy,” The Journal of Finance, 23(4), 589—609.

BAE, K., AND V. GOYAL (2009): “Creditor Rights, Enforcement, and Bank Loans,”
The Journal of Finance, 64(2).

BEBCHUK, L., AND A. COHEN (2003): “Firms’ Decisions Where to Incorporate,”
The Journal of Law and Economics, 46(2), 383—425.

BEBCHUK, L., A. COHEN, AND A. F ERRELL (2009): “What matters in corporate
governance?,” Review of Financial Studies.

BENMELECH, E., M. GARMAISE, AND T. Mosxowrrz (2005): “Do Liquidation

Values Affect Financial Contracts? Evidence from Commercial Loan Contracts and
Zoning Regulat10n*,” The Quarterly Journal of Economics, 120(3), 1121—1154.

BOLTON, R, AND D. SCHARFSTEIN (1996): “Optimal debt structure and the number
of creditors,” The Journal of Political Economy, 104(1), 1—25.

BOOT, A., S. GREENBAUM, AND A. THAKOR (1993): “Reputation and discretion
in financial contracting,” The American Economic Review, 83(5), 1165—1183.

CHAVA, S., D. LIVDAN, AND A. PURNANANDAM (2009): “Do shareholder rights
affect the cost of bank loans?,” Review of Financial Studies.

CHAVA, S., AND M. ROBERTS (2008): “How does ﬁnancing impact investment? The
role of debt covenants,” Journal of Finance, 63, 2085—2121.

CHEN, X., J. HARFORD, AND K. L1 (2007): “Monitoring: Which institutions mat-
ter?,” Journal of Financial Economics, 86(2), 279—305.

COLES, J ., N. DANIEL, AND L. NAVEEN (2006): “Managerial incentives and risk-
taking,” Journal of Financial Economics, 79(2), 431—468.

CORE, J ., AND W. GUAY (1999): “The use of equity grants to manage optimal
equity incentive levels,” Journal of Accounting and Economics, 28(2), 151—184.

 

(2002): “Estimating the value of employee stock option portfolios and their
sensitivities to price and volatility,” Journal of Accounting Research, pp. 613—630.

CORE, J ., W. GUAY, AND R. VERRECCHIA (2003): “Price versus non-price perfor-
mance measures in optimal CEO compensation contracts,” The Accounting Review,

78(4), 957—981.
128

CREMERS, K., AND V. NAIR (2005): “Governance mechanisms and equity prices,”
The Journal of Finance, 60(6), 2859—2894.

CREMERS, M., V. NAIR, AND C. WEI (2004): “The impact of shareholder control
on bondholders,” Yale School of Management and N YU.

DENNIS, S., AND D. MULLINEAUX (2000): “Syndicated loans,” Journal of ﬁnancial
intermediation, 9(4), 404—426.

DIAMOND, D. (2004): “Presidential address, committing to commit: Short-term debt
when enforcement is costly,” The Journal of Finance, 59(4), 1447—1479.

EDMANS, A., X. GABArx, AND A. LANDIER (2008): “A Multiplicative Model of
Optimal CEO Incentives in Market Equilibrium,” Review of Financial Studies.

EISDORFER, A. (2008): “Empirical evidence of risk shifting in ﬁnancially distressed
ﬁrms,” Journal of Finance, 63(2), 609-637.

EISFELDT, A., AND A. RAMPINI (2006): “Capital reallocation and liquidity,” Jour-
nal of Monetary Economics, 53(3), 369—399.

FAULKENDER, M., AND M. PETERSEN (2006): “Does the source of capital affect
capital structure?,” Review of Financial Studies, 19(1), 45.

FOCARELLI, D., A. POZZOLO, AND L. CASOLARO (2007): “The pricing effect of
certiﬁcation on syndicated loans,” Journal of Monetary Economics.

GAVAZZA, A. (2010): “Asset liquidity and ﬁnancial contracts: Evidence from aircraft
leases,” Journal of Financial Economics, 95(1), 62-84.

GIROUD, X., AND H. MUELLER (2010): “Corporate governance, product market
competition, and equity prices,” .

GOMPERS, P., J. ISHII, AND A. METRICK (2003): “Corporate Governance and
Equity Prices,” Quarterly Journal of Economics, 118(1), 107—155.

GOPALAN, R., V. NANDA, V. YERRAMILLI, AND O. DRIVE (2009): “Lead arranger
reputation and the loan syndication market,” .

GRULLON, (3., AND R. MICHAELY (2005): “Corporate payout policy and product
market competition,” .

HANLON, M., S. RAJGOPAL, AND T. SHEVLIN (2003): “Are executive stock options

associated with future earnings?,” Journal of Accounting and Economics, 36(1-3),
3—43.

HART, 0., AND J. MOORE (1994): “A theory of debt based on the inalienability of
human capital,” The Quarterly Journal of Economics, 109(4), 841—879.

IVASHINA, V. (2009): “Asymmetric information effects on loan spreads,” Journal of
Financial Economics, 92(2), 300—319. '

JENSEN, M., AND W. MECKLING (1976): “Theory of the ﬁrm: Managerial behavior,

agency costs and ownership structure,” Journal of ﬁnancial economics, 3(4), 305—
360.

JONES, J ., W. LANG, AND P. NIGRO (2005): “Agent bank behavior in bank loan

syndications,” Journal of Financial Research, 28(3), 385.
129

KLOCK, M., S. MANSI, AND W. MAXWELL (2005): “Does Corporate Governance
Matter to Bondholders?,” Journal of Financial and Quantitative Analysis, 40(04),
693—719.

LEE, S., AND D. MULLINEAUX (2004): “Monitoring, ﬁnancial distress, and the
structure of commercial lending syndicates,” Financial management, pp. 107—130.

LEVINE, C., AND J. HUGHES (2005): “Management compensation and earnings-
based covenants as signaling devices in credit markets,” Journal of Corporate F i-

nance, 11(5), 832—850.

ORTIZ-MOLINA, H., AND G. PHILLIPS (2009): “Asset Liquidity and the Cost of
Capital,” .

PETERSEN, M. (2009): “Estimating standard errors in ﬁnance panel data sets: Com-
paring approaches,” Review of Financial Studies, 22(1), 435.

QIAN, J., AND P. STRAHAN (2007): “How Laws and Institutions Shape Financial
Contracts: The Case of Bank Loans,” The Journal of Finance, 62(6), 2803—2834.

RAJAN, R., AND A. WINTON (1995): “Covenants and collateral as incentives to
monitor,” The Journal of Finance, 50(4), 1113—1146.

SHLEIFER, A., AND R. VISHNY (1992): “Liquidation values and debt capacity: A
market equilibrium approach,” Journal of Finance, pp. 1343—1366.

SUFI, A. (2007): “Information asymmetry and ﬁnancing arrangements: Evidence
from syndicated loans,” The Journal of Finance, 62(2), 629.

VASVARI, F. (2009): “Equity compensation and the pricing of syndicated loans,” .

WALD, J ., AND M. LONG (2007): “The effect of state laws on capital structure,”
Journal of Financial Economics, 83(2), 297—319.

130

CHAPTER 3

The Effects of Asset Liquidity on.

the Loan Pricing

Abstract

I examine the impact of firms’ asset liquidation value which is proxied by asset liquid-
ity on debt contracting using comprehensive US syndicated loan contracts. I employ
various measures of the asset liquidity at the three—digit industry level. The borrower,
belonging to a higher asset liquidity industry at the time of loan origination, experi-
ences a lower loan spread. However as the maturity of a loan increases the impact of
the asset liquidity declines. Furthermore I found that the loan Spreads are affected
by the number of lenders because of the concern about coordination failures, and by
the possibility of the borrower’s being involved in M&A activities. These results are
consistent with incomplete contracting and transaction cost theories of liquidation

value and ﬁnancial structure.

131

3. 1 Introduction

Although there are extensive theoretical predictions on how much important the
liquidation values of borrower’s assets are on debt contracts 1, there are a few empirical
studies including reCent ones such as Benmelech, Garmaise, and Moskowitz (2005),
Gavazza (2010).

In a ﬁnancing contracting, the borrower and the lender try to make an optimal
contract given the information available at the time of contracting. Among many
considerations which are important for determining the terms of the loan, the infor-
mation on the liquidation values of debtor’ assets, which will be seized up when the
debtor is not able to repay (Aghion and Bolton, 1992, Bolton and Scharfstein, 1996,
Hart and Moore, 1994) is not readily available because the liquidation values depend
on the market condition at the time of liquidation of the assets.

If we believe that despite the lack of complete information, the two parties will
make the most of the then available information on the future liquidation values as
rational economic agents behave. The information that they are interested in can be
found simply from the previous one if they employ adaptive updating method for the
information. As a proxy for the measure of the liquidation values during the span of
loan maturity, I will use three-digit industry asset liquidity variables of the prior year
to loan origination (more detailed explanation will be found in the next chapter).

One of the questions on the effects of the liquidation values on loan structures, such
as interest rates, maturity, and loan size is to know whether higher liquidation values
contribute to lowering interest rates as predicted by theories. For this question, Ben-
melech, Garmaise, and Moskowitz (2005), Gavazza (2010) demonstrate that higher
liquidation value(or higher asset liquidity) lowers interest rates(for leasing contract-
ing, markups of operating lease rates). These studies focus on US airplane sector. In

my paper, I want to cover whole US industries other than the ﬁnancial sector. This

 

1see, Aghion and Bolton (1992), Diamond (2004), Hart and Moore (1994)
132

broader coverage of industries and more detailed loan contracts using comprehensive
ﬁrm loan data have some advantages. First, the results that will be drawn will be
a general answer to whether the liquidation values of ﬁrrns’ assets affect the loan
contracts. Second, I can control ﬁrm and loan characteristics that if missed, may
affect the answer. Third, because of the rich details in loan terms, I can explore other
aspects of the question.

The rest of the paper is organized as follows. Section II discusses how to construct
proxies for the liquidation values and data sources. Section III describes data sum-
mary and model speciﬁcation. Section IV presents the empirical results and discuss
some relevant points. Section V discusses model speciﬁcation and presents the results

of a robust check regression, and Section VI concludes.

3.2 ’ Measures of Asset Liquidity

I use asset liquidity as a proxy for the liquidation values for ﬁrms’ assets. The liquid-
ity of assets can be deﬁned as the extent to which asset sellers sell them at the prices
that are the closest to their intrinsic values of use. The market liquidity of an asset
becomes high when there is a high volume of transactions in an industry where the as-
set is traded. To measure the volume of transactions, I construct various indices using
two different data sources, Compustat and SDC. As in Eisfeldt and Rampini (2006),
I also deﬁne the liquidity of ﬁrm’s assets as the ratio of the sum of ﬁrms’ acquisitions
and sales of property, plant and equipment over ﬁrms’ total assets(or Property, Plant
and Equipment, or Capital Expenditures) in an industry. Firms are grouped on the
basis of their three-digit SIC. This measure represents the intensity of used capital
transactions on the industry-year basis and asset liquidity. Although the items for ac-
quisitions(AQC in Compustat) include various instances that are important in terms

of accounting practices, they may not capture the ‘pure’ transactions which are more

133

relevant for the purpose of this paper. As a complementary measure, I construct a
ratio which is the sum of the values of transactions of M&A, and asset sell/ purchase
to corresponding industry’s total assets using the SDC data as in Ortiz-Molina and
Phillips (2009). The SDC comprehensively registers all the transactions that entail
M&A and asset acquisitions. To standardize the industry level liquidity, the total
value of transactions in the three-digit SIC industry is divided by the industry’s total
assets. In doing so while making sure the two data sets are compatible, I exclude the
transactions when the sellers or purchasers are not identiﬁed based on the Compustat

data.

3.3 Sample Construction, Empirical Speciﬁcation

3.3.1 Sample Construction

My data come from the Compustat Database, the Securities Data Corporation (SDC),
and the Dealscan of Dealogic. Loan data from Dealscan were withdrawn by the month
of June in 2007. Firms, the borrowers in the loans, are identiﬁed as Chava and Roberts
(2008) did 2. I exclude ﬁnancial ﬁrms (SIC codes 6000 to 6999) and loans contracted
before 19903.

3.3.2 Empirical Speciﬁcation and Key Variables

To test the hypothesis, the following basic testing model is speciﬁed
Spreadijt = a + 7t + SLiquidityJ-t + ngtci + Eijt (3.1)

Spreadijt is the All-in-drawn spread on the bankloan measured over the LIBOR.

“n is added to control macroeconomic common shocks to the ﬁrms. Liquidityjt rep-

 

2I thank for their generosity in providing ﬁrm identiﬁcation information
3Previous researchers found the Dealscan data is not comprehensive before 1990. Many papers
also use the loans originated after 1990 134

resents industry-level liquidity measures. These various measures based on (Eisfeldt
and Rampini, 2006, Ortiz-Molina and Phillips, 2009) are included one by one for each
regression. Xijt consists of ﬁrm’s and loan’s characteristics. Since the errors are
believed to be correlated in the industry level, the equation (1.1) is estimated while
taking account of clustering effects in the industry level (Petersen, 2009).

In the analysis, I include control variables that have a high explanatory power
for the interest spread. logAT is the logarithm of ﬁrm’s book-value total assets as
a measure of ﬁrm’s size. Larger ﬁrms tend to be established earlier so that they
have better access to external ﬁnancing. Bl is book leverage((DLC + DLTT)/AT)
constructed by dividing added values of long-term and short—term debts over total
assets. Higher leveraged ﬁrms are more likely to default. A ﬁrm’s proﬁts also is a
very important indicator for assessing the probability of loan repayments. Proﬁt is
constructed by ﬁrm’s EBITDA over total assets(OIBDP / AT). Tang is ﬁrm’s tangible
asset ratio against total assets (PPENT/ AT). One of the concerns that the lenders
hold in lending decision is how to recover defaulted loans. The recovery depends on

4 is constructed as in Altman (1968) to

the value of debtor’s tangible assets; Zscore
capture the ﬁrm’s probability of default and its lower score translates into a higher
cost of external ﬁnancing.

Along with ﬁrm characteristic variables, I include some loan-speciﬁc variables such
as loan maturity, facility size, number of lenders, and dummy for whether the loan is
secured. Longer maturity loans are more likely to be exposed to the uncertainty of
default in the lender’s standpoint of view. The facility size is related with the spread
because the amount of risk exposure is dependent on the size of loss. The number of
lenders is also controlled because in the lender’s point of view, when the loan goes

wrong, the more lenders means lesser burden that the individual lender should take.

I also control for macroeconomic effects by including year dummies for the years of

 

43.3 * PI + SALE + 1.4 * RE + 1.2 * (ACT-LCT))/AT, where PI is Pretax Income; SALE is

Net Sale; RE is Retained Earnings; ACT is Current Assets, and LCT is Current Liabilities
135

loan origination.

3.3.3 Summary Statistics

Table 3.1 provides year-wise summary of loan and asset liquidity data. The data
set includes 22,433 loans, and covers 235 unique three-digit SIC industries. Average
size of tranche is 260.7 million dollars with maturity being 3 years. On average, 7.5
lenders join a loan syndicate. The Spread over LIBOR that the borrowers pay is
185.7 basis points. Among the entire sample loans of which 75% are ‘Secured’ in that
when the borrowers go default, the lenders can claim on ﬁrms’ assets to recover their
loans. A number of data observations in the Dealscan does not report whether the
loans are ‘Secured’, for such reason, I lose about 7000 observations in the regression
analysis. However in lender’s point of view, whether the loans are ‘Secured’ or not
is very important when it comes to loan pricing because secured assets at least stem
losses for the lenders when the borrowers cannot make repayments. For this reason,
the variable, ‘Secured’ should be included in the empirical models although I lose
many observations.

The average size of ﬁrms in the sample is 6.4 billion dollars in terms of their book-
value total assets. These average ﬁrms are comparatively large because of the sample
selection. Although the Dealscan covers comprehensively US syndicated loans, for the
study, the ones that can be identiﬁed by the robust matching rules are only included
in the ﬁnal sample and in general the ﬁrms in the Compustat are large and established
earlier. The sample mean of Property, Plant, and Equipment(PPENT) is 2.7 billion
dollars and of which 7.4% change ownership by selling and purchasing. On average,
the ﬁrms’ leverage is 0.33 and proﬁt ratio stands at 0.14. The ﬁrms hold 35% of total
assets as tangible assets, Property, Plant, and Equipment. The ﬁnancial condition

of the average ﬁrm in terms of its bankruptcy probability is Grey: Altman (1968)

136

grouped three categories according to ﬁrm’s zscore5, Distress, Grey, and Safe when
its zscore falls below 1.8, or between 1.8 and 2.99, and above 2.99 respectively.

As measures of asset liquidity, I have used six different deﬁnitions. Reall is the
sum of SPPE(Sales of Property, Plant, and Equipment) and AQC(Acquisition of
assets). Reall/Capital investment is deﬁned as the ratio of Reall over capital in-
vestment(property, plant, and equipment - capital expenditures) that in the sample
is 66%. Instead Reall, AQC and SPPE are divided by capital investment respec-
tively, AQC/ Capital investment is seven times larger than the ratio of SPPE/ Capital
investment.

I also construct measures in that AQC, and SPPE are divided by the corresponding
industry’s one-year lagged AT, and PPENT. Furthermore, to see the effects of the
change in the composition of the asset liquidity, the ratio of SPPE/Reall is added as
another measure. In line with Ortiz-Molina and Phillips (2009), Shleifer and Vishny
(1992), asset liquidity is determined by the number of competing ﬁrms in the same
industry which are able to purchase used capital. In the sample, NoPotBuy is deﬁned
as the number of rival ﬁrms in the three-digit SIC industry that have debt ratings. On
average there are 20 competing ﬁrms in an industry, however the standard deviation
is very large. Another index for asset liquidity for robust checks is constructed using
the SDC data set in the same manner for the previous measures. Compared with
Reall/lag( A T) sourced from the Compustat, Total transactions/A ssetsﬁfrom the SDC

is slightly smaller.

 

5The zscore is an indicator for the probability of bankruptcy of a ﬁrm in two years
6Total transactions are the values of trades, and when this information is missing, the value of
zero is assigned

137

3.4 Empirical Results

In this section I present main empirical ﬁndings’on the effect of asset liquidity on
loan pricing. I divide the discussion into sub—sections. The ﬁrst one is to examine
the question of whether asset liquidity affects the loan pricing, and if it does, then
to what extent. In the second subsection, I examine when loan maturity becomes
longer, which is understood as increased uncertainty on asset liquidity, whether the
impact of asset liquidity is still effective. In the following subsection, I will investigate

how coordination failure among the lenders affects the loan pricing.

3.4.1 Main results

I begin by examining how the three-digit industry level asset liquidity affects the
loan pricing while comprehensively controlling ﬁrm and loan characteristics in the
regressions. Table 3.2 presents the results of the regressions with the various asset
liquidity measures. The dependent variable, All in spread drawn(henceforth aisd) is
taken in logarithm to control for skewness in the data7.

As can be seen from Column (1), the coefﬁcient of Reall/ CAPX V is negative and
signiﬁcant. Since this is a linear model, the coefﬁcient is also equal to the marginal
effect. To estimate the economic signiﬁcance of this result, we examine a predicted
decline in loan spreads that would happen if the asset liquidity of industry increases by
one standard deviation. Based on the estimates from the model (1) of Table 3.2, one
standard deviation increase causes the spread to decline by 43pr, and this amounts
to 2.5% decline around the average spread that the borrowers pay additionally over
LIBOR. 1

Such effects of the asset liquidity are also observed following the various asset liquid-

 

7'Results are qualitatively unchanged if we directly use All in spread drawn as the dependent

variable
80.76*0.03*187.82)

138

ity measures: AQC/ CAPX V, Reall/lag(AT), AQC/lag(AT), and Reall/lag(PPEN T),
although their statistical signiﬁcances vary. Unlike those measures, the coefﬁcient of
the SPPE/Reall is shown positive, which also squares with my prediction. In the used
capital market, if the number of sellers is larger than that of buyers (in my sample,
which is represented as the difference in the composition structure of reallocation).
When the SPPE/Reall increases, in other words, the ﬁrms that belong to an industry
whose SPPE/Reall gets larger relatively compared with that of other industries, be-
cause of lower probability for the ﬁrms to sell their used capital in the capital market
at their intrinsic values, the lenders ask higher interest rates to avoid the loss from
selling borrower’s assets at lower prices in the future. The coefﬁcients on the loan’s
and ﬁrm’s characteristics are not only highly signiﬁcant, save Number of Lenders, but
also Show right signs. As the ﬁrm’s size, proﬁtability, and tangibility increase, the
spread declines, however when the leverage, a measure of ﬁnancial condition of the
ﬁrm, worsens, the spread increases.

Table 3.3 reports the estimates of the regressions in which the measures of asset
liquidity are interacted with dummies constructed based on the Zscore. diss indicates
that the ﬁrm’s zscore is below 1.8, and when the score falls between 1.8 and 2.99, the
ﬁrm is categorized as grey. safe is assigned when the score is above 2.99. As shown,
only the coefﬁcient of the interaction of the asset liquidity with grey is statistically
signiﬁcant at the 1% signiﬁcance level. Not only that, in the economic sense, when the
increase in the Reall/Capri} is made by one standard deviation the spread declines by
11bp , which is more than double the decrease in the spread in Column (1) of Table
3.2. At the time of lending decision, the most obscure ﬁrm to lenders in terms of its
bankruptcy probability in the future will be the one whose zscore is in the grey area.
Although basing their lending decision on the information that is currently available
at the time of lending, the lenders cannot be sure of the future event regarding the

grey ﬁrms. If we believe lenders act rationally, the lenders will be much less concerned

139

with the liquidation value of a firrn’s assets when the industry to which the borrowing
ﬁrm belongs is more liquid. This lightens lenders’ concern and will be expressed as

more deep cuts in the spread, especially for the grey ﬁrms.

3.4.2 Maturity and Uncertainty

As maturity gets longer, lenders will be exposed to uncertainty because of the current
information becomes less informative in the future. As we have observed, the cur—
rent information on asset liquidity affects the loan pricing at the reciprocal manner.
However as maturity increases, the impact of asset liquidity declines. This predic-
tion is shown more obviously in Table 3.4 when as asset liquidity measures I use
Reall/Caprv and AQC/Caprv. However for the measures such as Reall/lag(At) and
Reall/lag(PPEN T), the same effect is not observed. Such difference can be found in
the inherent characteristic of capital expenditure. Capital expenditure is one of the
most volatile aggregate expenditures in the macroeconomic sense. Because of such
high volatility, the information on it is more likely to become obsolete as the time at
which the information is used moves farther away from the current time.

In contrast, the effect of the change in the composition of asset liquidity,
SPPE/Reall, turns into being positive as maturity increases. If lenders have to liqui—
date the borrower’s assets in the narrow span of time, they have to make a deep cut in
selling prices; however when loan maturity is longer, the incident of deep cutting will
be expected to be less severe compared with the case in the current situation. This
aspect appeared as a relatively slower speed of spread increase for the loan whose

maturity is longer.

3.4.3 Coordination failure and Pricing

At the time when lenders have to liquidate, the liquidation value of assets can be

expressed as a function of not only markﬁoliquidity, but also the number of lenders.

As the number of lenders increase, coordination among lenders will be more difﬁcult
in deciding how to dispose of the borrower’ assets. Delayed liquidation decision
may miss the optimal selling time and it translates into lower recovery values. To
compensate for the loss gap stemming from coordination failure, lenders will ask
higher spread. To check whether this claim is plausible in the contracting, I interact
liquidity measures with a dummy variable whose value is one when there is a sole
lender. Table 3.5 presents the results of the regressions. In model (1), and (2), as
asset liquidity increases, the spread declines as observed in the previous regressions.
The coefﬁcients of the interactions of Reall/Cava * Sole lender, and AQC/Caprv
I" Sole lender are negative and signiﬁcant at the 5% level. For alternative liquidity

measures, the prediction still holds.

3.5 Discussion and Robust test

In this section, I want to discuss issues of model speciﬁcation and robust test. The
effects of asset liquidity on the loan pricing may not be one dimension. The asset
liquidity can be understood as the degree of threat of being a target company for
risk-increasing M&A. If it is the case, then the speciﬁcation of the model should take
account of this aspect.

The results that we have seen so far support the theoretical prediction. In the
environment of higher asset liquidity, lenders ask lower spread for the ﬁrms whose
industry liquidity is higher. However, high asset liquidity does not necessarily work
for the direction of lowering the spread. When an industry experiences a high M&A
activity, which can increase asset liquidity to a larger extent, borrowing ﬁrms also can
be exposed to a risk for being a M&A target, or may actively involve themselves in
acquiring other ﬁrms. In either case whether the borrower is a target, or purchaser,

those transactions will heighten ﬁnancial risks to the given lenders because of bor—

141

rower’s increased leverage, or uncertainty of whether debt seniority is respected after
being acquiredg.

This different aspect of the effect of asset liquidity on loan pricing renders me
to think of the model speciﬁcation to capture this aspect. The model, which is
not only simple, but which also presents the two different regimes of the effects of
asset liquidity, is a polynomial function of second-degree in that the measures of asset
liquidity are squared. Although increased asset liquidity contributes to the alleviation
of lender’s concern for liquidation values, when the liquidity rises beyond a certain
level, the increased liquidity may adversely affect the loan pricing.

Table 3.6 reports the estimates of the regressions. In column (1), until Reall/Cava
reaches 2.25(0.09/(2*0.02)) which stands at the 95 percentile of the variable, the
negative relationship between the loan spread and asset liquidity holds as before.
For the other variables of the measure of asset liquidity, the points from which the
curves turn direction are also located at around the 95 percentile of each variable’s
distribution.

The concern for borrower’s being exposed to M&A target is legitimate. As ? has
observed, when the borrower steps up its takeover defense by increasing the number
of anti-takeover provisions, the borrower beneﬁts lower spread from external bank
ﬁnancing. In my paper, I examine the relationship between loan pricing and M&A
threat. To measure the intensity of M&A threat, I use NoPotBuyers, the number
of rival ﬁrms in the three-digit industry that have rating, and in the other model, I
interact the NoPotBuyers with the Zscore to compare the effect of M&A threat by
differing levels of borrower’s ﬁnancial condition.

Table 3.7 shows the results of the two regressions when the variable in interest is
NoPotBuyers alone, or its interaction with Zscore. In the column (1), the coefficient

of NoPotBuyers is not signiﬁcant at any level of signiﬁcance. However in column

 

9? examined how year-level M&A activity affects the loan pricing

142

(2), for the borrower whose ﬁnancial condition in the Safe zone, as the number of
rival ﬁrms, NoPotBuyers, gets larger, the loan spread increases. This results shed a
light on how lenders respond to the uncertainty of loan repayment: by increasing loan
spread when the probability of the borrowers’ being involved in M&A is high.

For the measures of asset liquidity, I have used accounting data from the Compustat.
In using this data, there may be an argument that because this is accounting data,
it may not represent the real asset sales and purchases. To support my ﬁndings, I
have constructed a very similar measure using the SDC data set. Since this data
set captures only real asset transactions, it may rightly represent the asset liquidity.
Instead of the previous liquidity measure, I estimate the coefﬁcient of the variable in
question while controlling the same ﬁrm and loan characteristics. The result shows the
same conclusion although its standard variance is larger compared with Reall/Caprv

so that the coefﬁcient is statistically signiﬁcant at the 10% levell.0

3.6 Conclusion

In the debt contracting theory, the liquidation values of borrower’s assets are of
importance because to force the borrower to repay loans, the lenders should be able
to trigger a credible liquidation threat. However when the triggering is constrained
by the market condition in which the lenders can not ﬁnd the second best user of the
assets, the lenders hesitant to enter into a debt contracting. This paper examined
whether increased asset liquidity results in lowering spread of the loans. Along with
the strong results in answering the question, the paper found the lenders are concerned
about the borrower whose default probability prediction is difﬁcult. Furthermore,
in relation with loan maturity, as the maturity gets longer, the effect of current-

period asset liquidity is shown to decline because of the increased uncertainty of asset

 

1Upon request, I will provide data and the result

143

liquidity in the future.

In addition, when more than one lender are participating in a loan syndicate,
coordination failures may occur and this affects the loan pricing. This effect is also
empirically observed in the paper. Finally, I explored the effects of excess asset.
liquidity on the loan pricing, and found that the spread increases when ﬁrm’s asset

is too liquid.

144

Table 3.1 Summary Statistics

 

 

Variable Mean STD. N
Number of Lenders . 7.5 8.9 22428
Tranche Amount(\$M) 260.7 619.7 22433
Maturity 1333.5 769.4 22433
All in spread Drawn(bp) 185.7 131.0 22433
Secured[0 or I] 0.75 0.43 15362
Assets(\SM) 6435.2 4781 1.0 22431
Property, Plant, and Equipment (\SM, PPEN": 2687.6 24524.8 22367
Sales of Assets(\$M, SPPE) 41.7 464.0 19810
Aquisitions of Assets(\$M, AQC) 153.1 1342.5 22087
Leverage 0.33 0.23 22366
Proﬁtability 0.14 0.10 22380
Tangibility 0.35 0.24 22363
Zscore l .86 1. l 8 20900
Reall/Capital investment 0.66 0.76 22433
AQC/Capital investment ' 0.58 0.75 22433
SPPE/Capital investment 0.08 0. 12 22433
Reall/lag(Assets) 0.04 0.04 22433
AQC/lag(Assets) 0.03 0.04 22433
Reall/lag(PPENT) ' 0.15 0.18 22433
SPPE/Reall 0.23 0.24 22433
NoPotBuy 20.4 26.9 22428
Total transactions/Assets(from SDC) 0.02 0.06 22431

 

145

Table 3.2 Effects of Asset Liquidity on the Loan Pricing

This table reports estimate results from the regressions which relate the effects of asset liquidity
to the loan pricing. The measures of asset liquidity are constructed at the three-digit industry
level. The explanatory variables include ﬁrm-speciﬁc, and loan-speciﬁc characteristics, ln(total
assets) refers to the book-value total assets of the ﬁrm in millions of USD before the loan
initialization. leverage is the total debt (long term plus short term) divided by total assets of the
ﬁrm. proﬁtability is the ratio of EBITDA of the ﬁrm to total assets of the ﬁrm. tangibility is the
ratio of Property, Plant, and Equipment over total assets of the ﬁrm. zscore is constructed
according to Altman (1968), ln(maturity) indicates the logarithm of the maturity of loan in days.
ln(facility size) is the logarithm of the size of facility in dollars. The number of lenders in the
syndicate is denoted by Number of Lenders. Secured is a dummy indicating whether the loan is
secured. Robust standard errors are estimated with clustered errors at the three-digit industry
level and are reported in parentheses. Regressions also include year ﬁxed effects, and loan
10%, 5% and 1%,

purpose indicators. ’, "', "* indicate the level of signiﬁcance at

 

 

respectively.
(1) (2) (3) (4) (5) (6)
Reall/CAPXV 003*"
(0.01)
AQC/CAPXV -0.03""'
(0.01)
Reall/lag(AT) -0.33*
(0.18)
AQC/lag(AT) -0.32"'
(0.18)
Reall/lag(PPENT) -0. 14*"
(0.04)
SPPE/Reall 0.09"
(0.04)
Ln(Facility size) -0.09"“'"' -0.09"”'”" -0.09"”‘”" -0.09*” -0.09“”" 009"”
(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)
Ln(Maturity) -0.02*" -0.02""" -0.02"'* -0.02""“ -0.02*" -0.02“
(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)
Number of Lenders 0.00 0.00 0.00 0.00 0.00 0.00
0.00 0.00 0.00 0.00 0.00 0.00
Secured [0 or 1] 0.72"" 0.72“" 0.72"" 0.72“" 0.72““ 072*"
(0.02) (0.02) (0.02) (0.02) (0.02) (0.02)
Ln(Total assets) -0.08"”'”" -0.08*“ 008"" -0.08"* -0.08"" -0.08*"
(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)
Leverage 051*" 0.51"” 0.51“" 051*" 0.50"" 0.51""
(0.03) (0.03) (0.03) (0.03) (0.03) (0.03)
Proﬁtability -0.85"* -0.84"* -0.84‘”'”" -0.84*" -0.84"'"”" 0.83""
(0.12) (0.12) (0.12) (0.12) (0.12) (0.12)
Tangibility -0. 13"" -0.13"”"‘ -0.10“ -0.1 I” 014*" -0.12"”'”"
(0.05) (0.05) (0.05) (0.04) (0.05) (0.04)
Zscore -0.03"”'”" -0.03**"' -0.03"”"' -0.03"”""' -0.03"”'”" -0.03*"
(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)
Intercept 6.76“” 676*“ 675*” 6.76"" 6.78“" 6.74""
(0.10) (0.10) (0.09) (0.09) (0.10) (0.10)
N 14325 14325 14325 14325 14325 14325
R squared 0.60 0.60 0.60 0.60 0.60 0.60

 

146

Table 3.3 Default Probability and Asset Liquidity

This table reports estimate results from the regressions which relate the effects of asset liquidity to
the loan pricing. The measures of asset liquidity are constructed at the three-digit industry level.
distress is a dummy whose value is one when zscore is below 1.8, grey is a dummy whose value is
one when zscore falls between 1.8 and 2.99, and when the zscore is above 2.99, a dummy, safe is
assigned one. The explanatory variables include ﬁrm-speciﬁc, and loan-speciﬁc characteristics,
ln(total assets) refers to the book-value total assets of the ﬁrm in millions of USD before the loan
initialization. leverage is the total debt (long term plus short term) divided by total assets of the
ﬁrm. proﬁtability is the ratio of EBITDA of the ﬁrm to total assets of the ﬁrm. tangibility is the
ratio of Property, Plant, and Equipment over total assets of the ﬁrm. zscore is constructed
according to Altman (1968), ln(maturity) indicates the logarithm of the maturity of loan in days.
ln(facility size) is the logarithm of the size of facility in dollars. The number of lenders in the
syndicate is denoted by Number of Lenders. Secured is a dummy indicating whether the loan is
secured. Robust standard errors are estimated with clustered errors at the three-digit industry level
and are reported in parentheses. Regressions also include year ﬁxed effects, and loan purpose
indicators. *, "', "* indicate the level of signiﬁcance at 10%, 5% and 1%, respectively.

(1) (2) (3) (4) (5) (6)
Reall/CAPXV "' distress -0.01
(0.01)
Reall/CAPXV " grey -0.08*"
(0.01)
Reall/CAPXV "' safe -0.03
(0.02)
AQC/CAPXV "‘ distress -0.01
(0.01)
AQC/CAPXV "‘ grey -0.08***
(0.02)
AQC/CAPXV * safe -0.03
(0.02)

Reall/lag(AT) * distress -0.01
(0.17)
Reall/lag(AT) * grey -1.31"*
(0.28)
Reall/lag(AT) "' safe -0.42
(0.41)
AQC/lag(AT) "‘ distress 0.04
(0.16)
AQC/lag(AT) " grey -1.38"""*
(0.29)
AQC/lag(AT) "' safe -0.45
(0.42)
Reall/lag(PPENT) "' distress -0.07
7 (0.05)
Reall/lag(PPENT) "‘ grey 029*"
(0.06)
Reall/lag(PPENT) " safe -0. 12
(0.07)
SPPE/Reall " distress 0.10“
(0.05)

 

147

Table 3.3 Default Probability and Asset Liquidity(cont.)
SPPE/Reall " grey 0.03
- (0.04)
SPPE/Reall " safe 018“"
(0.05)
Ln(Facility size) 009‘" 009*" 009*" -0.09"”‘“‘I 009*" -0.09""
(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)
Ln(Maturity) -0.02"“" -0.02"‘* 002" -0.02" -0.02** -0.02"
(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)
Number of Lenders 0.00 0.00 0.00 0.00 0.00 0.00
0.00 0.00 0.00 0.00 0.00 0.00
Secured [0 or 1] 0.72““l 072*" 072"" 072*" 0.72“" 072*"
(0.02) (0.02) (0.02) (0.02) (0.02) (0.02)
Ln(Total assets) -0.08*""" -0.08""‘ -0.08"“""' 008*" 008*" -0.08*"
(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)
Leverage 050"” 0.50“" 0.50“" 0.50"“ 0.50"“ 0.50""
(0.03) (0.03) (0.03) (0.03) (0.03) (0.03)
Proﬁtability 084"" -0.83"”'”" -0.82""‘”" 082"" -0.83"* 083""
(0.12) (0.12) (0.12) (0.12) (0.12) (0.12)
Tangibility 013*" 013*" -0.11‘”" -0.11" -0.14"“'”" -0.13"*
(0.04) (0.04) (0.04) (0.04) (0.04) (0.04)
Zscore -0.02""'"" -0.02*" 002"" -0.02"”'”" 002*" 003""
(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)
Intercept 675"" 6.75"" 6.75"" 6.75"" 677*" 6.74""
(0.09) (0.09) (0.09) (0.09) (0.10) (0.10)
N 14325 14325 14325 14325 14325 14325
R sMed 0.61 0.61 0.60 0.60 0.60 0.60

 

148

Table 3.4 Maturity and the Loan Pricing

This table reports estimate results from the regressions which relate the effects of asset liquidity
to the loan pricing. The measures of asset liquidity are constructed at the three-digit industry
level. The explanatory variables include ﬁrm-speciﬁc, and loan-speciﬁc characteristics, ln(total
assets) refers to the book-value total assets of the ﬁrm in millions of USD before the loan
initialization. leverage is the total debt (long term plus short term) divided by total assets of the
ﬁrm. proﬁtability is the ratio of EBITDA of the ﬁrm to total assets of the ﬁrm. tangibility is the
ratio of Property, Plant, and Equipment over total assets of the ﬁrm. zscore is constructed
according to Altman (1968), ln(maturity) indicates the logarithm of the maturity of loan in days.
ln(facility size) is the logarithm of the size of facility in dollars. The number of lenders in the
syndicate is denoted by Number of Lenders. Secured is a. dummy indicating whether the loan is
secured. Robust standard errors are estimated with clustered errors at the three-digit industry
level and are reported in parentheses. Regressions also include year ﬁxed effects, and loan
purpose indicators. *, "', *** indicate the level of signiﬁcance at 10%, 5% and 1%, respectively.

 

 

(1) (2) (3) (4) (5) (6)
Reall/CAPXV -0.18***
(0.07)
Reall/CAPXV " Ln(Maturity‘ 0.02"
(0.01)
AQC/CAPXV -0.19"'*"'
(0.07)
AQC/CAPXV * Ln(Maturity) 0.02"
(0.01)
RealVlag(AT) - l .30
(1.45)
RealVlag(AT) "' Ln(Maturity) 0.14
(0.20)
AQC/lag(AT) -l .98
(1.46)
AQC/lag(AT) "' Ln(Maturity) 0.23
(0.20)
RealVlag(PPENT) -0.46
(0.29)
RealVlag(PPENT) "' Ln(Maturity) 0.04
(0.04)
SPPE/Reall 0.94“"
(0.27)
SPPE/Reall * Ln(Maturity) 012*"
(0.04)
Ln(Facility size) -0.09*** -0.09"* -0.09"'** -0.09"”'"" -0.09"* 009‘"
(0.01) (0.01) (0.01) (0.01) (0.01) . (0.01)
Ln(Maturity) -0.03*** -0.03*** -0.03" -0.03*"”" -0.03** 0.01
(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)
Number of Lenders 0.00 0.00 0.00 0.00 0.00 0.00
0.00 0.00 0.00 0.00 0.00 0.00
Secured [0 or I] 072*" 072*" 0.72"" 0.72““‘ 072*" 0.72““l
(0.02) (0.02) (0.02) (0.02) (0.02) (0.02)
Ln(Total assets) -0.08"“'“" -0.08"'" -0.08"“'“" -0.08"* -0.08*"“" 008""
(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)

149

Leverage

Table 3.4 Maturity and the Loan Pricing(cont.)

051*” 051*" 0.51‘" 051*" 0.50“” 050*"
(0.03) (0.03) (0.03) (0.03) (0.03) (0.03)
Proﬁtability 084“" -0.84"'"‘"‘ -0.84*" 084"“ -0.84"‘" 083"“
(0.12) (0.12) (0.12) (0.12) (0.12) (0.12)
Tangibility 013*" -0. l3"”‘”‘I -0.10** -0.11"“" -0. 14*" -0.12’””“
(0.05) (0.05) (0.05) (0.05) (0.05) (0.04)
Zscore -0.03"“'“'I -0.03*** ~0.03"”"'l -0.03*** 003*" -0.03"”'”"
(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)
Intercept 686*" 686‘” 679*" 6.81"" 6.83“” 6.55”"
(0.11) (0.11) (0.10) (0.11) (0.12) (0.10)
N 14325 14325 14325 14325 14325 14325
R squared 0.60 0.60 0.60 0.60 0.60 0.60

 

150

Table 3.5 Coordination Failure and the Loan Pricing

This table reports estimate results from the regressions which relate the effects of asset liquidity
to the loan pricing. The measures of asset liquidity are constructed at the three-digit industry
level. The explanatory variables include ﬁrm-speciﬁc, and loan-speciﬁc characteristics, ln(total
assets) refers to the book-value total assets of the ﬁrm in millions of USD before the loan
initialization. leverage is the total debt (long term plus short term) divided by total assets of the
ﬁrm. proﬁtability is the ratio of EBITDA of the ﬁrm to total assets of the ﬁrm. tangibility is the
ratio of Property, Plant, and Equipment over total assets of the ﬁrm. zscore is constructed
according to Altman (1968), ln(maturity) indicates the logarithm of the maturity of loan in days.
ln(facility size) is the logarithm of the size of facility in dollars. The number of lenders in the
syndicate is denoted by Number of Lenders. Secured is a dummy indicating whether the loan is
secured. Robust standard errors are estimated with clustered errors at the three-digit industry
level and are reported in parentheses. Regressions also include year ﬁxed effects, and loan
purpose indicators. *, "', *" indicate the level of signiﬁcance at 10%, 5% and 1%,
respectively.

 

(1) (2) (3) (4) (5) (6)

 

Reall/CAPXV -0.02*
(0.01)
Reall/CAPXV " Sole lender -0.04"“"
(0.02)
AQC/CAPXV -0.02"”"
(0.01)
AQC/CAPXV "' Sole lender -0.04"
' (0.02)
RealVlag(AT) -0. 16
(0.18)
RealVlag(AT) * Sole lender -0.63***
(0.23)
AQC/Iag(AT) -0. 15
(0.18)
AQC/lag(AT) * Sole lender ~0.66""'”"
(0.25)
RealVlag(PPENT) -0.1 I"
(0.04)
RealVlag(PPENT) "' Sole lender -0.09
(0.07)
SPPE/Reall 0.09‘l
(0.05)
SPPE/Reall "' Sole lender 0.00
(0.06)
Ln(Facility size) -0.09"* 009*" -0.09"”"* -0.09*” -0.09”* -0.09*"
(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)
Ln(Maturity) -0.02" -0.02** -0.02"'""" -0.02"”"" -0.02"“" -0.02"”"
' (0.01) (0.01) (0.01) (0.01) (0.01) (0.01)
Number of Lenders 0.00 0.00 0.00 0.00 0.00 0.00
0.00 0.00 0.00 0.00 0.00 0.00
Secured [0 or 1] 072*" 072*" 0.72“" 0.72‘" 072*" 072*"
(0.02) (0.02) (0.02) (0.02) (0.02) (0.02)
Ln(Total assets) 008"" -0.08"* -0.08""'"" -0.08*"”" -0.08"""" -0.08"'"

(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)

151

Table 3.5 Coordination Failure and the Loan Pricing(cont.)

Leverage 051*" 051*" 050"" 050*" 0.50"" 051*"
(0.03) (0.03) (0.03) (0.03) (0.03) (0.03)
Proﬁtability -0.85"‘"* -0.85*"“" -0.85*** -0.85"'*" -0.85*" -0.83"”'”"
(0.12) (0.12) (0.12) (0.12) (0.12) (0.12)
Tangibility -0.12*"”" -0.12‘”'”" -0.10""" -0.10""" -0.14*** -0.l2‘”'”"
(0.05) (0.05) (0.05) (0.04) (0.05) (0.04)
Zscore -0.03*"”" -0.03*"‘* -0.03*** -0.03"”'”" -0.03"* -0.03"*
(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)
Intercept 684*" 683*" 683*" 683*" 682*" 673*"
(0.10) (0.10) (0.10) (0.10) (0.10) (0.10)
N 14325 14325 14325 14325 14325 14325
R squared 0.60 0.60 0.60 0.60 0.60 0.60

 

152

Table 3.6 Leverage Increasing Activity and the LoanPricing

This table reports estimate results from the regressions which relate the effects of asset liquidity to
the loan pricing. The measures of asset liquidity are constructed at the three-digit industry level. The
explanatory variables include ﬁrm-speciﬁc, and loan-speciﬁc characteristics, ln(total assets) refers to
the book-value total assets of the ﬁrm in millions of USD before the loan initialization. leverage is
the total debt (long term plus short term) divided by total assets of the ﬁrm. proﬁtability is the ratio
of EBITDA of the ﬁrm to total assets of the ﬁrm. tangibility is the ratio of Property, Plant, and
Equipment over total assets of the ﬁrm. zscore is constructed according to Altman (1968),
ln(maturity) indicates the logarithm of the maturity of loan in days. ln(facility size) is the logarithm
of the size of facility in dollars. The number of lenders in the syndicate is denoted by Number of
Lenders. Secured is a dummy indicating whether the loan is secured. Robust standard errors are
estimated with clustered errors at the three-digit industry level and are reported in parentheses.
Regressions also include year ﬁxed effects, and loan purpose indicators. *, “', *** indicate the level
of signiﬁcance at 10%, 5% and 1%, respectively.

 

(1) (2) (3) (4) (5) (6)

Reall/CAPXV 009*"
(0.02)
Reall/CAPXV2 002*"
(0.00)
AQC/CAPXV 009*"
(0.02)
AQC/CAva2 0.02***
(0.00)
RealVlag(AT) -0.96*"
(0.34)
RealVlag(AT)2 3.03"
(1.23)
AQC/lag(AT) -1.08***
(0.34)
AQC/lag(AT)2 3.77***
(1.31)
RealVlag(PPENT) -0.28**
(0.11)
RealVlag(PPENT)2 0.19
(0.15)
SPPE/Reall 0.07
(0.10)
SPPE/Reallz 0.02
(0.12)
Ln(Facility size) 009*" -0.09*** 009*" 009*" 009*** 009*"
(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)
Ln(Maturity) 002** 002** 002** 002** 002** 002"
(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)
Number of Lenders 0.00 0.00 0.00 0.00 0.00 0.00
0.00 0.00 0.00 0.00 0.00 0.00

153

Table 3.6 Leverage Increasing Activity and the Loan Pricing(cont.)

Secured [0 or 1] 0.72"" 072*" 0.72"“ 072‘" 0.72"" 0.72""
(0.02) (0.02) (0.02) (0.02) (0.02) (0.02)
Ln(Total assets) 008“" -0.08“”" 008"" -0.08"" -0.08""’ -0.08""
(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)
Leverage 0.51“” 051*" 0.51“” 051*" 050*" 0.51”"
(0.03) (0.03) (0.03) (0.03) (0.03) (0.03)
Proﬁtability -0.84""‘ 084*” -0.84"“"“ -0.83""‘ -0.84*** -0.83"”"*
(0.12) (0.12) (0.12) (0.12) (0.12) (0.12)
Tangibility -0. 14”" -0.15"* -0.1 I“ -0.11"”“ -0.15”* -0.12""
(0.05) (0.04) (0.04) (0.04) (0.04) (0.04)
Zscore -0.03*" -0.03*** -0.03"""* -0.03*" -0.03""'”' -0.03"”"*
(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)
Intercept 6.79"* 6.78"" 6.77"”Ml 6.77"" 6.79"" 6.74"“
(0.10) (0.10) (0.09) (0.09) (0.10) (0.10)
N 14325 14325 14325 14325 14325 14325
R squared 0.60 0.60 0.60 0.60 0.60 0.60

 

154

 

Table 3.7 Takevover and the Loan Pricing

This table reports estimate results from the regressions which relate the effects of asset
liquidity to the loan pricing. lNoPotBuy is the number of rival ﬁrms in the three-digit
industry and in the logarithm. distress is a dummy whose value is one when zscore is below
1.8, grey is a dummy whose value is one when zscore falls between 1.8 and 2.99, and when
the zscore is above 2.99, a dummy, safe is assigned one. The explanatory variables include
ﬁrm-speciﬁc, and loan-speciﬁc characteristics, ln(total assets) refers to the book-value total
assets of the ﬁrm in millions of USD before the loan initialization. leverage is the total debt
(long term plus short term) divided by total assets of the ﬁrm. proﬁtability is the ratio of
EBITDA of the ﬁrm to total assets of the ﬁrm. tangibility is the ratio of Property, Plant, and
Equipment over total assets of the ﬁrm. zscore is constructed according to Altman (1968),
ln(maturity) indicates the logarithm of the maturity of loan in days. ln(facility size) is the
logarithm of the size of facility in dollars. The number of lenders in the syndicate is denoted
by Number of Lenders. Secured is a dummy indicating whether the loan is secured. Robust
standard errors are estimated with clustered errors at the three-digit industry level and are
reported in parentheses. Regressions also include year ﬁxed effects, and loan purpose
indicators. *, "', "'" indicate the level of signiﬁcance at 10%, 5% and 1%, respectively.

 

 

(1) (2)
lNoPotBuy 0.01
(0.01)
lNoPotBuy "' distress 0.01
(0.01)
lNoPotBuy "' grey -0.01
(0.01)
lNoPotBuy "‘ safe 003*"
(0.01)
Ln(Facility size) -0.09*** -0.09"”'”"
(0.01) (0.01)
Ln(Maturity) -0.02"“" -0.02"“"
(0.01) (0.01)
Number of Lenders 0.00 0.00
0.00 0.00
Secured [0 or 1] 072*" 0.72"”
(0.02) (0.02)
Ln(Total assets) -0.08**"‘ -0.08***
(0.01) (0.01)
Leverage 051*" 0.50""
(0.03) (0.03)
Proﬁtability -035“: 083""
(0.12) (0.12)
Tangibility -0.1 I" -0.12**
(0.05) (0.05)
Zscore 002*" -0.03***
(0.01) (0.01)
Intercept 6.72“" 6.74""
(0.10) (0.10)
N 14321 14321
R squared 0.60 0.60

 

155

3.7 Bibliography

AGHION, P., AND P. BOLTON (1992): “An incomplete contracts approach to ﬁnan-
cial contracting,” The Review of Economic Studies, 59(3), 473—494.

ALTMAN, E. (1968): “Financial ratios, discriminant analysis and the prediction of
corporate bankruptcy,” The Journal of Finance, 23(4), 589—609.

BAE, K., AND V. GOYAL (2009): “Creditor Rights, Enforcement, and Bank Loans,”
The Journal of Finance, 64(2).

BEBCHUK, L., AND A. COHEN (2003): “Firms’ Decisions Where to Incorporate,”
The Journal of Law and Economics, 46(2), 383—425.

BEBCHUK, L., A. COHEN, AND A. FERRELL (2009): “What matters in corporate
governance?,” Review of Financial Studies.

BENMELECH, E., M. GARMAISE, AND T. Mosxowr'rz (2005): “Do Liquidation

Values Affect Financial Contracts? Evidence from Commercial Loan Contracts and
Zoning Regulation*,” The Quarterly Journal of Economics, 120(3), 1121—1154.

BOLTON, P., AND D. SCHARFSTEIN (1996): “Optimal debt structure and the number
of creditors,” The Journal of Political Economy, 104(1), 1-25.

BOOT, A., S. GREENBAUM, AND A. THAKOR (1993): “Reputation and discretion
in ﬁnancial contracting,” The American Economic Review, 83(5), 1165—1183.

CHAVA, S., D. LIVDAN, AND A. PURNANANDAM (2009): “Do shareholder rights
affect the cost of bank loans?,” Review of Financial Studies.

CHAVA, S., AND M. ROBERTS (2008): “How does ﬁnancing impact investment? The
role of debt covenants,” Journal of Finance, 63, 2085—2121.

CHEN, X., J. HARFORD, AND K. L1 (2007): “Monitoring: Which institutions mat-
ter?,” Journal of Financial Economics, 86(2), 279—305.

COLES, J ., N. DANIEL, AND L. NAVEEN (2006): “Managerial incentives and risk-
taking,” Journal of Financial Economics, 79(2), 431-468.

CORE, J ., AND W. GUAY (1999): “The use of equity grants to manage optimal
equity incentive levels,” Journal of Accounting and Economics, 28(2), 151—184.

(2002): “Estimating the value of employee stock option portfolios and their
sensitivities to price and volatility,” Journal of Accounting Research, pp. 613—630.

 

CORE, J ., W. GUAY, AND R. VERRECCHIA (2003): “Price versus non-price perfor-
mance measures in optimal CEO compensation contracts,” The Accounting Review,

78(4), 957—981.
156

CREMERS, K., AND V. NAIR (2005): “Governance mechanisms and equity prices,”
The Journal of Finance, 60(6), 2859—2894.

CREMERS, M., V. NAIR, AND C. WEI (2004): “The impact of shareholder control
on bondholders,” Yale School of Management and N YU.

DENNIS, S., AND D. MULLINEAUX (2000): “Syndicated loans,” Journal of ﬁnancial
intermediation, 9(4), 404—426.

DIAMOND, D. (2004): “Presidential address, committing to commit: Short-term debt
when enforcement is costly,” The Journal of Finance, 59(4), 1447—1479.

EDMANS, A., X. GABAIx, AND A. LANDIER (2008): “A Multiplicative Model of
Optimal CEO Incentives in Market Equilibrium,” Review of Financial Studies.

EISDORFER, A. (2008): “Empirical evidence of risk shifting in ﬁnancially distressed
ﬁrms,” Journal of Finance, 63(2), 609—637.

EISFELDT, A., AND A. RAMPINI (2006): “Capital reallocation and liquidity,” Jour-
nal of Monetary Economics, 53(3), 369—399.

FAULKENDER, M., AND M. PETERSEN (2006): “Does the source of capital affect
capital structure‘?,” Review of Financial Studies, 19(1), 45.

FOCARELLI, D., A. POZZOLO, AND L. CASOLARO (2007): “The pricing effect of
certiﬁcation on syndicated loans,” Journal of Monetary Economics.

GAVAZZA, A. (2010): “Asset liquidity and ﬁnancial contracts: Evidence from aircraft
leases,” Journal of Financial Economics, 95(1), 62—84.

GIROUD, X., AND H. MUELLER (2010): “Corporate governance, product market
competition, and equity prices,” .

GOMPERS, P., J. ISHII, AND A. METRICK (2003): “Corporate Governance and
Equity Prices,” Quarterly Journal of Economics, 118(1), 107—155.

GOPALAN, R., V. NANDA, V. YERRAMILLI, AND O. DRIVE (2009): “Lead arranger
reputation and the loan syndication market,” .

GRULLON, C., AND R. MICHAELY (2005): “Corporate payout policy and product
market competition,” .

HANLON, M., S. RAJGOPAL, AND T. SHEVLIN (2003): “Are executive stock options
associated with future earnings?,” Journal of Accounting and Economics, 36(1—3),
3—43.

HART, 0., AND J. MOORE (1994): “A theory of debt based on the inalienability of
human capital,” The Quarterly Journal of Economics, 109(4), 841—879.

IVASHINA, V. (2009): “Asymmetric information effects on loan spreads,” Journal of
Financial Economics, 92(2), 300—319.

JENSEN, M., AND W. MECKLING (1976): “Theory of the ﬁrm: Managerial behavior,

agency costs and ownership structure,” Journal of ﬁnancial economics, 3(4), 305—
360.

JONES, J., W. LANG, AND P. NIGRO (2005): “Agent bank behavior in bank loan
syndications,” Journal of Financial Research, 28(3), 385.
157

KLOCK, M., S. MANSI, AND W. MAXWELL (2005): “Does Corporate Governance
Matter to Bondholders?,” Journal of Financial and Quantitative Analysis, 40(04),
693-719.

LEE, S., AND D. MULLINEAUX (2004): “Monitoring, financial distress, and the
structure of commercial lending syndicates,” Financial management, pp. 107—130.

LEVINE, C., AND J. HUGHES (2005): “Management compensation and earnings-

based covenants as signaling devices in credit markets,” Journal of Corporate F i-
nance, 11(5), 832—850.

ORTIZ-MOLINA, H., AND G. PHILLIPS (2009): “Asset Liquidity and the Cost of
Capital,” .

PETERSEN, M. (2009): “Estimating standard errors in ﬁnance panel data sets: Com-
paring approaches,” Review of Financial Studies, 22(1), 435.

QIAN, J., AND P. STRAHAN (2007): “How Laws and Institutions Shape Financial
Contracts: The Case of Bank Loans,” The Journal of Finance, 62(6), 2803—2834.

RAJAN, R., AND A. WINTON (1995): “Covenants and collateral as incentives to
monitor,” The Journal of Finance, 50(4), 1113—1146.

SHLEIFER, A., AND R. VISHNY (1992): “Liquidation values and debt capacity: A
market equilibrium approach,” Journal of Finance, pp. 1343—1366.

SUFI, A. (2007): “Information asymmetry and financing arrangements: EVidence
from syndicated loans,” The Journal of Finance, 62(2), 629.

VASVARI, F. (2009): “Equity compensation and the pricing of syndicated loans,” .

WALD, J ., AND M. LONG (2007): “The effect of state laws on capital structure,”
Journal of Financial Economics, 83(2), 297—319.

158