a r t i c l e
i n f o
Article history:
Received 27 April 2015
Received in revised form
28 September 2015
Accepted 5 October 2015
Available online 22 October 2015
JEL classication:
G32
G33
F34
K20
a b s t r a c t
I nd that institutional arrangements have an impact on the real economy by affecting rms
choice between private and public debt and the subsequent nancing costs. Using new
debt issued by rms in 26 non-US countries, I nd, after controlling for rm characteristics
predicted by debt agency and information asymmetry theories, that the level of nancial
market development, the efciency of bankruptcy procedure, the integrity and enforceability of laws, and the transparency of nancial information have signicant impacts not only
on rms debt choice and yield to maturity in domestic debt market, but also their issuance
choice in the international debt market.
2015 Elsevier B.V. All rights reserved.
Keywords:
Debt placement
Bond yield
Legal protection
Institutional arrangements
1. Introduction
Debt is a major source of capital for rms worldwide. The extant literature nds that rms debt placement choice (i.e.,
public vs. private debt) and debt nancing costs are related to characteristics that reect the rms credit quality, debt
agency problems, level of information asymmetry, and otation costs (Cantillo and Wright, 2000; Denis and Mihov, 2003;
Hadlock and James, 2002; Krishnaswami et al., 1999). The implicit assumption of these studies is that capital markets are
competitive and perfectly elastic; hence, debt nancing decisions reect the outcomes when a rm solves an optimization
problem based on its fundamentals (Baker, 2009). To the extent that external capital markets are not always fully efcient
and perfectly competitive in many countries, rm debt nancing patterns are likely to be affected by external constraints.
Using a sample of rms located in 26 non-US countries from 1995 to 2008, I examine the effects of external institutional
constraints on rm debt nancing, above and beyond the effects of rm fundamentals.
A few papers have examined the effect of institutional arrangements on rms capital structure (Booth et al., 2001;
Demirguc-Kunt and Maksimovic, 1999; Fan et al., 2012). The focus of these papers is the effect of institutional arrangements
on the aggregate amount of debt in a rm. This study contributes to the literature by examining rms new debt issues instead.
363
Specically, I examine whether and how institutional arrangements systematically inuence a rms choice between public
and private debt for new debt issuance.
I focus on the following institutional features: the level of economic and nancial development, the extent and strength of
creditors legal protection, and the availability and transparency of nancial information. I expect these institutional features
to affect rm debt nancing by altering the relative advantages and disadvantages of public and private debt. Three channels
can be at work. First, the level of nancial market development determines the availability of capital and the accessibility of
different capital markets; this is a supply-side effect that strongly inuences a rms debt choice and the cost of debt. Second,
the quality and strength of creditors legal protection determine how effectively debt contracts can be enforced when a debt
covenant violation occurs. They also determine the efciency of the bankruptcy process, which can affect creditors recovery
rates. These two effects can alter a creditors assessment of a rms ex-ante bankruptcy risk. Third, the availability and
transparency of corporate nancial information is directly related to the level of information asymmetry, which inuences
both debt placement choice and nancing costs.
Country-level institutional environment signicantly affects rms choice between public and private debt. Consistent
with the rst channel, I nd that smaller banking sectors lead to a higher likelihood of issuing private placement bonds. This
suggests that, when faced with a limited loan supply, rms without access to public debt will need to use private placement
bonds as an alternative source of capital. Similarly, a large domestic equity market is associated with a low likelihood
of public debt issuance, indicating a substitution effect between the two major public nancing sources. The results are
consistent with the ndings in Graham and Harvey (2001) and Leary (2009) that the supply of different forms of capital has
a signicant impact on rms capital structure and the mix of debt sources. The size of the insurance industry is positively
and signicantly associated with the likelihood of issuing public bonds, consistent with the idea that the development of the
insurance industry is an indicator of economic growth and nancial sector development (Beck and Webb, 2003; Fan et al.,
2012).
I also nd supportive evidence for the second channel. In countries with weak legal protection and enforcement, private
placement bonds that allow investors to actively monitor managers and limit manager discretion are more likely to be
issued. In contrast, when the legal system has strong control over managerial self-dealing and systematic corruption, the
likelihood of issuing new debt publicly is higher. Moreover, I nd that institutional arrangements in the domestic country
can affect the sensitivity of the debt issuance choice to rms information asymmetry. The choice of public versus private is
less affected by rm tangibility in cases where the external legal system provides stronger protection to creditors.
The third channel suggests that nancial transparency mitigates the information asymmetry between rms and outside
creditors, and makes it less costly for rms to borrow from the public debt market. From this perspective, I expect that
higher accounting standards and greater scope of credit information will raise the likelihood that a rm will issue public
debt. However, the results only partially support this prediction.
Issuing debt in international capital markets has become an increasingly common practice in the last decade, especially
after the creation of the euro and the subsequent rapid development of the Eurobond market. In fact, about 27 percent of
bonds in the sample are placed internationally. The evidence indicates that a rms ability to access to international debt
markets is affected by the issuers nationality. Firms from countries with highly developed capital markets, stronger creditor
protection, and high-quality nancial information are more likely to not only borrow in international debt markets but also
issue international debt publicly. In contrast, rms from developing countries more often issue domestic debt. Also, private
lending contracts are used more often in the international debt market when the issuers domestic country has weak legal
enforcement regarding managerial self-dealing.
Debt nancing costs, calculated as the yield to maturity minus risk-free interest rate at the date of issuance, are also
affected by the institutional environment. Higher GDP per capita, larger banking sector and equity market are associated with
lower bond yields. Better creditor rights protection and more transparent nancial information also reduce debt nancing
costs. Moreover, I nd that the yield of public bonds is more heavily affected by institutional environment than that of private
placement bonds.
This study is directly related to an emerging stream of literature exploring the supply-side effects of corporate nancing
and capital structure decisions. Greenwood and Vayanos (2013) and Graham et al. (2013) analyze the effect of monetary
and scal policies on corporate capital structure. Leary and Roberts (2005) and Stohs and Mauer (1996) test the effect of
interest rate on rms capital structure decisions. Leary (2009) studies the effect of bank loan supply shocks on rms capital
structure and the mix of debt sources. This study suggests that institutional arrangements can affect the availability of funds
in different capital markets and the investors preference for different types of debt contracts, and in turn affect individual
rms choice between public and private debt.
This study is also closely related to a large body of literature that explores the economic impact of institutional factors,
beginning with King and Levine (1993) and La Porta et al. (1997, 1998). Issues that have been studied include the effect
of legal institutions on nancial development (Djankov et al., 2008), corporate governance (Doidge et al., 2007), corporate
investment (Ellis et al., 2012), capital structure (Booth et al., 2001; Fan et al., 2012; Arosa et al., 2014), and access to nancing
(Aggarwal and Goodell, 2014). This study adds to this literature by showing that institutional arrangements have a real
impact on rm decisions when new debt is issued and that the external institutional environment signicantly affects the
choice between private and public debt, domestic versus international debt, and the nancing costs associated with the new
debt issuance.
364
1
According to the statistics published by Bank for International Settlements (BIS), at the end of 2009, the total outstanding amount of international
corporate bonds worldwide is about $3.05 trillion US dollars, and the total outstanding amount of domestic corporate bonds worldwide is $6.197 trillion
US dollars.
365
size of domestic bond market or when large amounts of foreign capital are needed. Nevertheless, we still control for the
aggregate bank credit to private sectors in order to account for the effect of bank loan market on the choice of public or
private placement bonds.
3.2. Data source and sample statistics
Information on new bond issuance is obtained from the SDC Global New Issuances database for the years 19952008.
Variable-rate bonds, convertible bonds, and bonds with equity features, such as bonds with warrants or rights, are eliminated,
leaving only xed-rate, straight bonds in the sample. Rule 144A-private placements are also excluded.2 Mortgage-backed
and asset-backed bonds are excluded from the sample due to their contract features that clearly differ from those of corporate
bonds. Bonds issued by nancial companies (SIC 60006999) are also excluded from the sample. Financial information on
bond issuers is primarily obtained from the SDC and WorldScope databases. I include countries in the sample only if they have
at least three private placements and three public bond offerings during the sample period. A total of 26 non-US countries
meet this requirement, including 8 developing countries and 18 developed countries. Geographically, the sample covers
countries from Asia, Europe, the Americas and the Pacic. A list of the sample countries is given in Table 1.
The sample includes 5984 new public bond issues, and 5398 of them are issued by rms in developed countries. The
sample has 1158 new private placement bond issues, of which 645 are issued by rms from developed countries. Table 2
Panel A shows the characteristics of public and private placement bonds. On average, public bonds have larger total proceeds
($265.59 million versus $104.28 million), longer term to maturity (85.47 months versus 60.94 months), and lower yield to
maturity (3.34% versus 5.22%) than private placement bonds. Issuers vary substantially across the two types of bonds. Public
bonds issuers are much larger than private bond issuers. In terms of credit ratings, almost all of the bonds included in the
sample have investment grade ratings. Of the public issuers, 13% are classied as high-tech rms, compared to 10% of the
private issuers. Public bond issuers also have a slightly higher level of tangible assets. The leverage ratio is similar across the
two groups.
3.3. Institutional arrangements
The size of a countrys banking sector relative to GDP and the size of equity market relative to GDP are obtained from
the WDI database annually. Table 2 Panel B shows that both banking sector and equity market are larger in developed
(common-law) countries than they are in developing (civil-law) countries. Insurance penetration is calculated as the value
of total insurance premium over GDP. I obtain this data from Fan et al. (2012). On average, the insurance industry is much
larger in developed countries than in developing countries.
The extent to which creditors have legal protection is measured by the Creditor Rights Index, constructed by La Porta
et al. (1998). The index ranges from zero (weak protection) to four (strong protection). Common-law countries have better
creditor rights protection, with an average index score of 3, while the average score for civil-law countries is 1.88. I measure the strength of law enforcement using control over managerial self-dealing, control over corruption, and prevalence
of tax evasion. The anti-self-dealing index (Djankov et al., 2008) is a numerical measure of the intensity of public and private enforcement of regulations on managerial self-dealing, and higher scores indicate stronger enforcement. On average,
common-law countries in the sample have higher scores than civil law countries. Corruption (i.e., the extent to which public
power is exercised to gain private benet) is negatively associated with a countrys legal enforcement. The control over
corruption index, which I obtain from the Worldwide Governance Indicators, ranges from 0.84 to 2.39 in the sample, with
higher scores representing stronger legal enforcement. On average, developed countries and common law countries have
better control over corruption. I include the tax evasion index of the World Economic Forum as the third measure of a
countrys legal enforcement level. This index ranges from 2.03 to 8.54 in the sample, with higher scores indicating higher
tax evasion and weaker tax law enforcement.
WDI provides an index that measures the rules affecting the scope, accessibility, and quality of credit information available
through public and private credit registries. I use this index to measure the availability of credit information. I also use La
Porta et al. (1998) Accounting Standard Index to measure the quality and comprehensiveness of nancial information for
each sample country. Developed (common law) countries on average have better quality credit and nancial information
than developing (civil law) countries.
4. Regression analysis
4.1. Debt nancing choice
Table 3 reports the estimates of logit regressions testing the effect of country-level institutional arrangements on rms
debt choices. A countrys stage of economic development has a signicant inuence. I nd that rms in developed countries
2
Fenn (2000) provides a detailed description of the unique characters of 144-A private placement bonds. Unlike traditional private placements, Rule
144A-private placements are structured to facilitate inter-institutional trading in the secondary market; thus, they tend to resemble public offering bonds
more than traditional, non-144A private bonds.
366
Table 1
List of sample countries.
Nation
Number of
public
bonds
Number of
private
placement
Aggregate size of
issuer over equity
market cap
Developed
country
OECD
member
G8
member
Legal
origins
Australia
Austria
Brazil
Canada
China
Finland
France
Germany
Hong Kong
India
Indonesia
Italy
Japan
Luxembourg
Malaysia
Netherlands
New Zealand
Peru
Philippines
Singapore
South Korea
Spain
Sweden
Switzerland
Thailand
United Kingdom
43
23
80
30
17
13
341
122
30
157
54
58
3825
19
9
144
5
117
3
46
127
27
63
41
149
441
36
14
14
51
44
4
37
21
19
144
17
5
229
12
196
35
6
3
28
80
35
3
14
11
67
33
11%
9%
14%
2%
2%
14%
47%
35%
6%
18%
10%
34%
37%
8%
11%
18%
7%
62%
20%
6%
87%
14%
22%
7%
50%
23%
1
1
0
1
0
1
1
1
1
0
0
1
1
1
0
1
1
0
0
1
1
1
1
1
0
1
1
1
0
1
0
1
1
1
0
0
0
1
1
1
0
1
1
0
0
0
1 (After 1996)
1
1
1
0
1
0
0
0
1
0
0
1
1
0
0
0
1
1
0
0
0
0
0
0
0
0
0
0
0
0
1
Common
Civil
Civil
Common
Civil
Civil
Civil
Civil
Common
Common
Civil
Civil
Civil
Civil
Common
Civil
Common
Civil
Civil
Common
Civil
Civil
Civil
Civil
Civil
Common
Nation
Banking
sector
size
Equity
market
size
Insurance
penetration
Credit
Rights
Index
Tax
evasion
Anti-Self
Dealing
Index
Control over
corruption
Depth of
credit
info
Accounting
Standard
Index
Australia
Austria
Brazil
Canada
China
Finland
France
Germany
Hong Kong
India
Indonesia
Italy
Japan
Luxembourg
Malaysia
Netherlands
New Zealand
Peru
Philippines
Singapore
South Korea
Spain
Sweden
Switzerland
Thailand
United Kingdom
108.47%
125.59%
75.67%
183.23%
131.10%
65.06%
109.29%
137.50%
140%
61.80%
50.06%
105.24%
297.92%
101.04%
149.33%
152.21%
122.15%
20.66%
60.99%
81.81%
89.69%
143.28%
105.77%
176.45%
134.26%
142.52%
99.42%
25.06%
40.74%
109.38%
127.56%
123.06%
75.01%
46.90%
348.68%
51.33%
26.64%
47.31%
73.66%
153.20%
149.17%
111.67%
33.93%
28.44%
41.80%
176.82%
49.14%
88.36%
109.41%
239.97%
59.66%
149.26%
4%
2%
0
3%
1%
7%
6%
3%
3%
1%
1%
3%
9%
5%
2%
5%
2%
0
1%
3%
9%
2%
4%
7%
1%
9%
3
3
1
1
2
1
0
3
4
2
3
2
3
n/a
3
3
4
0
1
3
3
2
1
1
3
4
5.78
5.69
2.89
6.93
3.81
7.38
4.54
4.6
6
2.21
2.73
3.13
6.36
7.63
6.09
5.84
7.29
n/a
2.03
8.54
4.63
4.82
3.97
6.84
3.74
6.69
0.76
0.21
0.27
0.64
0.76
0.45
0.38
0.28
0.96
0.58
0.65
0.42
0.5
0.28
0.95
0.21
0.95
0.45
0.21
1
0.47
0.37
0.33
0.26
0.81
0.95
1.99
1.83
0.0067
2.27
0.183
2.39
1.5
1.85
1.5
0.22
0.84
0.77
1.28
1.96
0.47
2.27
2.32
0.12
0.4
2
0.38
1.32
2.35
2.25
0.25
2.09
5
6
5
6
2
5
4
6
4
2
2
6
6
n/a
6
5
5
6
3
4
5
5
4
5
4
6
75
54
54
74
n/a
77
69
62
69
57
n/a
62
65
n/a
76
64
70
38
65
78
62
64
83
68
64
78
This table lists the 26 sample countries and their institutional arrangements. The number of public (private placement) bonds is the total number of public
(private placement) bonds issued by rms located in this country during the sample period. Variables are dened in Appendix 1. The average values of
annual variables are reported.
are more likely to access the public bond market. High GDP per capita is also associated with a higher likelihood of public
bond issuance. The development of nancial markets also clearly affects rms debt choice: rms are less likely to borrow
from the private placement bond market when the domestic banking sector is large, and rms are less likely to borrow
from the public bond market when the domestic stock market is large. Although Fan et al. (2012) do not nd a signicant
impact of insurance penetration on rms overall debt-to-asset ratio, I nd it to be signicantly and positively related to the
likelihood that new debt will be issued publicly.
367
Table 2
Difference of institutional arrangements across subsamples.
Public bond
Panel A
Proceeds ($ millions)
Maturity (months)
Yield to maturity at issuance date
Issuer total assets ($ millions)
Issuer leverage
Bonds with investment grade rating
% of high tech rms
Industry tangibility
Panel B
Number of public issuances
Number of private placements
Average GDP per capita ($)
Banking sector size
Equity market size
Insurance penetration
Credit Rights Index
Tax evasion
Anti-Self-Dealing Index
Control over corruption
Depth of credit information
Accounting Standard Index
Mean
Median
Mean
Median
265.59
85.47
3.34%
2264.93
0.45
99%
13%
0.48
127.5
72
2.36%
1081.69
0.44
N/A
N/A
0.42
104.28
60.94
5.22%
879.16
0.46
98%
10%
0.44
31.6
60
4.80%
103.39
0.43
N/A
N/A
0.42
Full sample
Developed countries
Developing countries
5984
1158
21,812.45
1.18
0.99
3.6%
2.24
5.21
0.54
1.18
4.68
66.43
5398
645
29,465.49
1.33
1.14
4.77%
2.41
5.93
0.52
1.80
5.12
69.06
616
513
5549.73
0.85
0.66
0.88%
1.88
3.36
0.59
-0.19
3.75
59.00
761
565
25,137.81
1.24
1.40
3.75%
3.00
6.19
0.85
1.55
4.75
72.13
5223
593
20,247.57
1.16
0.82
3.5%
1.88
4.74
0.41
1.02
4.65
63.40
Panel A reports the statistics of bond and issuer characteristics. Panel B reports the total number of new bond issues within each subgroup. The average
value of each institutional variable within each subgroup is also reported.
Table 3
Logit regression: institutions and choice between public and private bonds.
Model 1
Developed country indicator
Model 2
Model 3
Model 4
0.19***
(0.04)
0.27**
(0.07)
0.89***
(0.07)
0.29***
(0.02)
1.05***
(0.09)
0.59***
(0.09)
0.12
(0.08)
0.89***
(0.05)
0.21***
(0.06)
0.35***
(0.08)
1.87***
(0.07)
0.26***
(0.058)
0.04
(0.32)
0.36**
(0.12)
0.91***
(0.08)
Constant
0.29***
(0.06)
3.42***
(0.29)
5.54***
(0.70)
0.17***
(0.05)
8.73***
(0.65)
0.27
(0.38)
Yes
7142
0.11
Yes
7067
0.14
Yes
6947
0.18
Yes
6935
0.16
This table reports the results of logit regressions that estimate a rms likelihood of issuing public bonds using country level institutional arrangements.
The dependent variable equals one if the bond is issued publicly and zero if the bond is placed privately. All explanatory variables are dened in Appendix
1. Standard errors are clustered by country and year reported in parenthesis.
*
Signicance at 10% level.
**
Signicance at 5% level.
***
Signicance at 1% level.
368
Table 4
Logit regression: interaction of institutions and rm characteristics.
Model 1
Model 2
Model 3
Model 4
Model 5
1.59***
(0.07)
0.22
(0.20)
1.57***
(0.49)
0.78***
(0.19)
0.19***
(0.05)
0.58**
(0.34)
1.11***
(0.03)
1.12***
(0.06)
0.07
(0.10)
1.18**
(0.50)
0.37**
(0.16)
0.20***
(0.04)
0.49*
(0.28)
3.31***
(0.80)
1.11***
(0.06)
0.19
(0.12)
1.34***
(0.50)
0.41**
(0.16)
0.19***
(0.04)
0.7**
(0.28)
3.1***
(0.83)
1.09***
(0.06)
0.07
(0.10)
1.33***
(0.50)
0.6
(0.43)
0.20***
(0.04)
0.51*
(0.28)
0.9***
(0.27)
1.12***
(0.06)
0.11
(0.10)
1.69***
(0.51)
2.07*
(0.46)
0.20***
(0.04)
0.65**
(0.28)
1.06***
(0.28)
0.6***
(0.22)
1.12***
(0.21)
0.15
(0.17)
0.32***
(0.04)
0.09
(0.17)
2.56**
(1.07)
1.38***
(0.33)
0.81***
(0.14)
0.92
(0.14)
13.35***
(2.77)
0.71***
(0.08)
0.38***
(0.11)
1.22***
(0.13)
0.61***
(0.08)
0.11
(0.11)
0.89***
(0.14)
0.69***
(0.08)
0.39***
(0.11)
1.25***
(0.13)
0.58**
(0.08)
0.08
(0.11)
0.89***
(0.13)
Interaction terms
Industry tangibility * creditor rights protection
0.03
(0.16)
0.53***
(0.08)
1.44***
(0.77)
3.64***
(0.36)
0.87
(0.27)
3.31***
(1.20)
0.09
(0.15)
Constant
0.8
(2.48)
6.04***
(1.06)
7.72***
(1.01)
8.79***
(0.87)
2.87***
(0.79)
6.37***
(0.93)
Yes
5131
0.32
Yes
5299
0.26
Yes
5299
0.27
Yes
5299
0.26
Yes
5299
0.27
This table reports the results of logit regressions that estimate a rms likelihood of issuing public bonds. The dependent variable equals one if the bond is
issued publicly and zero if the bond is placed privately. All explanatory variables are dened in Appendix 1. Standard errors are clustered by country and
year reported in parenthesis.
*
Signicance at 10% level.
**
Signicance at 5% level.
***
Signicance at 1% level.
Stronger statutory creditor protection, measured by the Creditor Rights Index, signicantly increases a rms likelihood
of issuing public bonds. Furthermore, if a country has strong public and private enforcement of anti-self-dealing laws and
better control over corruption, its rms have a higher likelihood of issuing debt publicly. Lastly, weak legal enforcement,
measured by the prevalence of tax evasion, lowers a rms likelihood of issuing bonds publicly. Surprisingly, the results in
regression 4 of Table 3 suggest that more nancial transparency tends to lower the likelihood of issuing public bonds, which
contradicts my prediction.
I consider both rm-level and country-level factors in regression 1 of Table 4. Institutional arrangements remain significantly associated with rms debt choice after controlling for rm- and bond-specic characteristics. Furthermore, I nd
that rms that face higher debt agency costs benet more from stronger legal protection of creditor rights. A rms choice
369
Table 5
International bond issues.
Model 1
International = 1
Model 2
Public = 1
Model 3
Public = 1
0.96***
(0.15)
2.8***
(0.12)
0.7***
(0.10)
0.39***
(0.04)
0.37***
(0.07)
3.25***
(0.62)
2.4***
(0.19)
2.47***
(0.11)
1.26***
(0.10)
0.6
(2.49)
14.52***
(1.65)
0.44***
(0.17)
0.65***
(0.15)
0.19
(0.15)
0.17***
(0.03)
0.07
(0.06)
2.09***
(0.36)
0.87***
(0.17)
0.64***
(0.13)
0.35***
(0.14)
1.25***
(0.13)
0.07
(0.12)
0.09***
(0.03)
4.05**
(1.84)
0.01***
(0.08)
4.94***
(1.10)
4.71***
(1.22)
Yes
Yes
Yes
Yes
5131
0.48
Yes
1820
0.16
Yes
1820
0.14
This table tests a rms likelihood of issuing bond internationally. The dependent variable in model 1 equals one if the
bond is issued outside the country and zero if the bond is issued domestically. Models 2 and 3 test the likelihood of
issuing bonds publicly for the subsample of international bonds. The dependent variable equals one if this international
bond is issued publicly and zero if it is issue privately. Same variables that describe the bond and issuer characteristics
are also included in the regressions as in Table 4. All variables are dened in Appendix 1. Standard errors are clustered
by country and year, they are reported in parenthesis.
*
Signicance at 10% level.
**
Signicance at 5% level.
***
Signicance at 1% level.
between public and private bonds is less sensitive to its proportion of tangible assets and embedded growth options when
its domestic country has stronger creditor protection laws and effective enforcement. As reported in regressions 2 and 3
of Table 4, the interaction terms of industry tangibility with creditor rights protection and the anti-self-dealing index are
both negative and signicant, while the interaction terms of the high-tech indictor with creditor rights protection and the
anti-self-dealing index in regressions 4 and 5 are positive and signicant. The latter result suggests that high-tech rms have
better access to the public bond market when creditors have better legal protection.
4.2. International bond issues
Table 5 tests the effect of a countrys institutional characteristics on its domestic rms access to the international debt
market. Using a logit model, I nd that GDP per capita is positively associated with the likelihood of international bond
issuance. A large domestic banking sector is negatively associated with issuing international debt, while a well-developed
domestic equity market and insurance industry increase a rms likelihood of issuing debt internationally. Furthermore, I nd
that strong laws to protect creditor rights and control managerial self-dealing and corruption also have signicantly positive
association with rms likelihood of issuing bonds internationally. The prevalence of tax evasion, which is an indicator of
weak tax law enforcement, signicantly lowers a rms likelihood of issuing bonds internationally, whereas the availability
of reliable credit information signicantly increases it.
Models 2 and 3 of Table 5 present the results on the choice between public issuance and private placement for the
subsample of international bonds. A larger banking sector is associated with a lower likelihood of issuing publicly. Moreover,
greater legal constraints on managerial self-dealing and control over corruption signicantly increase the likelihood that
international debt will be issued publicly. Consistent with the predicted effect of nancial transparency, I nd that higher
nancial reporting standards largely increase a rms likelihood of issuing public bonds in the international debt market.
Compared with the results in Table 3, where the effects of reliable and high-quality credit and nancial information are
found to be negative, the evidence in Table 5 suggests that nancial transparency is given much more weight by foreign
investors than by domestic investors.
370
Table 6
Bond yield and institutional arrangements.
Public bond
Lambda
Log (GDP per capita)
Model 1
Model 2
Model 2
Model 3
Model 4
0.7***
(0.11)
0.41***
(0.07)
0.58***
(0.05)
0.98***
(0.11)
0.47***
(0.07)
1.93***
(0.03)
0.93***
(0.12)
0.43***
(0.07)
0.18
(0.16)
1.16***
(0.09)
1.73***
(0.30)
0.35
(0.07)
1.63***
(0.04)
2.74***
(0.14)
Developed indicator
Banking sector size
Equity market size
1.73***
(0.04)
0.07
(0.05)
0.2***
(0.02)
0.47***
(0.03)
0.16***
(0.046)
2.84***
(0.18)
0.49***
(0.05)
Yes
Yes
Yes
Yes
Yes
No
Yes
5203
0.73
No
Yes
5229
0.61
No
Yes
5229
0.62
No
Yes
5229
0.48
No
Yes
5229
0.63
This table reports the estimates on bond yield, adjusted for self-selection by including the Lambda estimated in the rst step. The econometric model is as
follows:
Dit = 1 if Xit + Wit + it > 0
(1)
(2)
(3)
where D is an indicator variable and is equal to one if the rm chooses to issue a public bond and zero for a private placement bond. The coefcient d
captures the average yield spread of public bonds over private bonds. X is a vector of bond characteristics, W is a matrix of institutional factors, and 0 and
1 are the inverse Mills ratios calculated from the rst-step probit regression. The dependent variable is bond yield, measured by yield minus risk-free
interest rate. Public bond indicator equals one if the bond is issued publicly. Institutional variables are included and dened in Appendix 1. The variables
that describe the bond and issuer characteristics are also included in the regressions as in Table 4, although they are not reported here to save space.
Standard errors are clustered by country and year reported in parenthesis.
*
Signicance at 10% level.
**
Signicance at 5% level.
***
Signicance at 1% level.
371
Regression 5 of Table 6 tests the effect of nancial transparency on yield spread. The interaction between the public bond
indicator and the depth of credit information is negative and signicant, while the public bond indicator itself is positive and
signicant. This indicates that, when it is difcult to obtain credit information through publicly available channels, public
bonds have a higher yield than similar private placement bonds, suggesting that the private debt market is likely to provide
cheaper nancing than, and hence has a comparative advantage over, the public bond market when nancial transparency
is poor. Conversely, public bonds become signicantly cheaper than private placement bonds when nancial information
becomes more transparent.
4.4. Robustness tests
To ensure the results are not heavily inuenced by a single country, I re-estimate all regressions excluding Japanese
rms, which account for a large proportion of the sample. All results stay qualitatively the same. Moreover, I also consider
the effect of cyclical patterns of security issuance by controlling for hot or cold market conditions based on the aggregate
amount of bond issuance, the results are not affected by these additional controls. These robustness results are available
upon request.
5. Conclusion
Both theoretical and empirical studies have shown that rm-level factors drive a rms choice between public and private
debt as well as the cost of debt nancing. The results in this study indicate that the institutional environment of a rms home
country also signicantly affects these decisions. I nd that economic and nancial development, creditor rights protection
and legal enforcement, and nancial transparency all signicantly alter a rms choice between private and public bonds in
domestic and international bond markets, as well as the corresponding bond yield of the specic debt instrument they choose
to issue. The results are robust after controlling for various rm-level characteristics and the domestic and international debt
market conditions.
This study suggests that corporate nancing decisions are likely to be jointly determined by an internal effect and an
external effect, such as the constraints existing in capital markets. I have shown that key institutional factors have a large
impact on individual rms debt nancing, perhaps through their impact on shaping the overall efciency of the domestic
capital markets and the inherent risks associated with the strength of the legal system.
Appendix 1. Data denition
Date item
Bond characteristics
Proceeds
Maturity
Investment grade rating
Firm characteristics
Firm size
Firm leverage
High tech rm
Industry tangibility
Institutional variables
Developed countries
GDP per capita
Common law
Civil law
Banking sector size
Equity market size
Insurance penetration
Creditor Rights Index
Denition
Data source
SDC
SDC
SDC
Natural logarithm of a rms total book assets at the end of scal year before bond
issuance
(Long term debt + short term debt)/total assets
Based on AeAs denition of high tech industries
The median value of tangible ratio for all rms in the same industry based on
four-digit SIC code
WorldScope
WorldScope
WorldScope and AeA
WorldScope
World Bank
WDI
LLSV 1998
LLSV 1998
WDI
WDI
Fan et al. (2012)
372
Appendix 1 (Continued)
Date item
Denition
Data source
(3) Secured creditors are ranked rst in the distribution of the proceeds that
result from the disposition of the assets of a bankrupt rm.
(4) The debtor does not retain the administration of its property pending the
resolution of the reorganization.
Anti-self-dealing index
Control over corruption
Tax evasion
Depth of credit information
Accounting Standard Index
References
Aggarwal, R., Goodell, J.W., 2014. Cross-national differences in access to nance: inuence of culture and institutional environments. Res. Int. Bus. Financ.
31, 193211.
Arosa, C.M.V., Richie, N., Schuhmann, P.W., 2014. The impact of culture on market timing in capital structure choices. Res. Int. Bus. Financ. 31, 178192.
Baker, M., 2009. Capital market-driven corporate nance. Annu. Rev. Financ. Econ. 1, 181205.
Beck, T., Demirguc-Kunt, A., Levine, R., 2000. A new database on the structure and development of the nancial sector. World Bank Econ. Rev. 14, 597605.
Beck, T., Webb, I., 2003. Economic, demographic, and institutional determinants of life insurance consumption across countries. World Bank Econ. Rev. 17,
5188.
Blackwell, D.W., Kidwell, D.S., 1988. An investigation of cost differences between public sales and private placements of debt. J. Financ. Econ. 22, 253278.
Booth, L., Avivazian, V., Demirguc-Kunt, A., Maksimovic, V., 2001. Capital structures in developing countries. J. Financ. 56, 87130.
Campa, J.M., Kedia, S., 2002. Explaining the diversication discount. J. Financ. 57, 17311762.
Cantillo, M., Wright, J., 2000. How do rms choose their lenders? An empirical investigation. Rev. Financ. Stud. 13, 155189.
Carey, M.S., Prowse, S., Rea, J., Udell, G., 1994. The Economics of the Private Placement Market (Board of Governors of the Federal Reserve System (U.S.)).
Chemmanur, T.J., Fulghieri, P., 1994. Reputation, renegotiation, and the choice between bank loans and publicly traded debt. Rev. Financ. Stud. 7, 475506.
Demirguc-Kunt, A., Maksimovic, V., 1999. Institutions, nancial markets, and rm debt maturity. J. Financ. Econ. 54, 295336.
Denis, D.J., Mihov, V.T., 2003. The choice among bank debt, non-bank private debt, and public debt: evidence from new corporate borrowings. J. Financ.
Econ. 70, 328.
Diamond, D.W., 1984. Financial intermediation and delegated monitoring. Rev. Econ. Stud. 51, 393414.
Diamond, D.W., 1991. Monitoring and reputation: the choice between bank loans and directly placed debt. J. Polit. Econ. 99, 689721.
Djankov, S., La Porta, R., Lopez-de-Silanes, F., Shleifer, A., 2008. The law and economics of self-dealing. J. Financ. Econ. 88, 430465.
Doidge, C., Karolyi, G.A., Stulz, R.M., 2007. Why do countries matter so much for corporate governance? J. Financ. Econ. 86, 139.
Ellis, J.A., Moeller, S.B., Schlingemann, F.P., Stulz, R.M., 2012. Globalization, country governance, and corporate investment decisions: an analysis of
cross-border acquisitions. SSRN eLibrary.
Fan, J.P.H., Titman, S., Twite, G., 2012. An international comparison of capital structure and debt maturity choices. J. Financ. Quant. Anal. 47, 2356.
Fenn, G.W., 2000. Speed of issuance and the adequacy of disclosure in the 144A high-yield debt market. J. Financ. Econ. 56, 383405.
Galai, D., Masulis, R.W., 1976. The option pricing model and the risk factor of stock. J. Financ. Econ. 3, 5381.
Graham, J.R., Harvey, C.R., 2001. The theory and practice of corporate nance: evidence from the eld. J. Financ. Econ. 60, 187243.
Graham, J.R., Leary, M., Roberts, M., 2013. Financial Crowding Out, working paper.
Greenwood, R., Vayanos, D., 2013. Bond supply and excess bond returns, working paper.
Hadlock, C.J., James, C.M., 2002. Do banks provide nancial slack? J. Financ. 57, 13831419.
Hearn, B., 2014. Institutional impact on the expropriation of private benets of control in North America. Res. Int. Bus. Financ. 30, 123.
Jensen, M.C., Meckling, W.H., 1976. Theory of the rm: managerial behavior, agency costs and ownership structure. J. Financ. Econ. 3, 305360.
King, R.G., Levine, R., 1993. Finance and growth: Schumpeter might be right. Q. J. Econ. 108, 717737.
Krishnaswami, S., Spindt, P.A., Subramaniam, V., 1999. Information asymmetry, monitoring, and the placement structure of corporate debt. J. Financ.
Econ. 51, 407434.
La Porta, R., Lopez-De-Silanes, F., Shleifer, A., Vishny, R., 1997. Legal determinants of external nance. J. Financ. 52, 11311150.
La Porta, R., Lopez-De-Silanes, F., Shleifer, A., Vishny, R., 1998. Law and nance. J. Polit. Econ. 106, 11131155.
Leary, M.T., Roberts, M.R., 2005. Do rms rebalance their capital structures? J. Financ. 60, 25752619.
Leary, M.T., 2009. Bank loan supply, lender choice, and corporate capital structure. J. Financ. 64, 11431185.
Leland, H.E., Pyle, D.H., 1977. Informational asymmetries, nancial structure, and nancial intermediation. J. Financ. 32, 371387.
Li, K., Prabhala, N.R., 20052007. Self-selection models in corporate nance. In: Eckbo, B.E. (Ed.), Handbook of Corporate Finance: Empirical Corporate
Finance, vol. 1. Elsevier/North-Holland Handbook of Finance Series, pp. 3950 (Chapter 2).
Myers, S.C., 1977. Determinants of corporate borrowing. J. Financ. Econ. 5, 147175.
Ramakrishnan, R.T.S., Thakor, A.V., 1984. Information reliability and a theory of nancial intermediation. Rev. Econ. Stud. 51, 415432.
Stohs, M.H., Mauer, D.C., 1996. The determinants of corporate debt maturity structure. J. Bus. 69, 279312.