Information Asymmetry and Capital

Structure Around the World

Wenlian Gao and Feifei Zhu∗

This Version: February 2013

∗ Authors’ contact information: Gao, [email protected], Chinese Academy of Finance

and Development, Central University of Finance and Economics. Zhu, [email protected], Hawai‘i
Pacific University. We appreciate the helpful comments of Lilian Ng, Bohui Zhang, Jung Chul
Park, and Yawen Jiao as well as the seminar participants in 2011 Financial Management As-
sociation Meeting, University of Wisconsin - Milwaukee, the Central University of Finance and
Economics, and 2012 China International Conference in Finance for their helpful comments and
suggestions.

Electronic copy available at: http://ssrn.com/abstract=2126451

 Information Asymmetry and Capital Structure
Around the World

ABSTRACT

This paper examines the relation between information asymmetry, capital structure and
the cost of capital across countries, particularly focusing on how the relation is influenced
by the various aspects of institutional environment. Results show that firms with more
informational asymmetries tend to use more debt capital but less long-term debt possibly
because of the differential impact of information asymmetry on the cost of different types
of capital. Furthermore, the positive association between information asymmetry and
market leverage is more pronounced in countries with developed banking sectors, and less
prominent in common-law countries or countries with more extensive disclosure practice.

Keywords: Information Asymmetry, Capital Structure, Cost of Capital, Country Char-

acteristics, IFRS
JEL Classification Number: G12, G14, G32

Electronic copy available at: http://ssrn.com/abstract=2126451

 1

1 Introduction

In this study, we examine the interplay of information asymmetry, capital structure, and
cost of capital, the three major components of a firm’s financial structure across countries
and assess the extent to which it is influenced by economic environment. The role of
information asymmetry in firms’ financial decision-making can be traced back to the
seminal work of Myers and Majluf (1984). Under their pecking order theory, equity capital
is the most information-sensitive security whereas debt capital is much less sensitive to
the adverse selection problem. Consequently, outside investors would require a higher
adverse selection risk premium on equity than on debt and firms with intense information
asymmetry would borrow more and end up with high leverage.1 On the other hand,
the literature suggests that certain country characteristics can facilitate or hinder firms’
use of a particular security as they affect transaction costs, stakeholders’ legal rights
enforcement, and the hazards of raising external capital etc.(See Rajan and Zingales
(1995), Booth et al. (2001), Giannetti (2003), Claessens, Djankov, and Mody (2001),
Djankov, Hart, McLiesh, and Shleifer (2008), and La Porta, Lopez-de-Silanes, Shleifer,
and Vishny (1997), (1998)) More recently, Fan, Titman, and Twite (2010) find that
country level institutional factors explain a significant portion of the variation in leverage
and debt maturity structures. Thus, we conjecture that the impacts of informational
asymmetries on firms’ financing decisions could be affected by institutional arrangements.
Specifically, we are interested in whether the effect of information asymmetry on capital
structure differs across countries and how certain country variables explain the difference.

Our sample consists of 90,514 firm-year observations for 13,019 industrial firms from
1
In the past few decades, a large body of work has been devoted to scrutinize whether pecking order
accurately describes the observed financing behavior. One strand of studies finds evidence in support of
the pecking order theory and argue that it is a good approximation of reality for some instances (See
Shyam-Sunder and Myers (1999) and Fama and French (2002)). The other strand of studies shows that
the pecking order theory fails under certain conditions (See Jung, Kim, and Stulz (1996), Frank and
Goyal (2003), Bolton and Dewatripont (2005), and Leary and Roberts (2010)). Recently, several studies
further attempt to evaluate the core assumption of pecking order theory, i.e., information asymmetry as
a determinant of capital structure decisions (See Bharath, Pasquariello, and Wu (2009) and Agarwal and
O’Hara (2007)).

Electronic copy available at: http://ssrn.com/abstract=2126451

 2

39 developed and developing countries in the period of 1997 to 2007. We construct mea-
sures of capital structure, the cost of capital, and firm-level controls based on accounting
information from Worldscope and stock trading information from Datastream. We em-
ploy several liquidity measures developed by the market microstructure literature to proxy
for the information environment surrounding a firm. Though the market-based measures
are originally developed to capture the information asymmetry among different investor
groups, they can characterize the information advantage held by firm insiders and the
resulting adverse selection costs (Bharath, Pasquariello, and Wu (2009)). To capture the
common variability in these measures, we form an index of adverse selection as the first
principle component of the correlation matrix for the adverse selection proxies which are
cross-sectionally standardized in each country year.2

We first show that our information asymmetry measure explains the cross-sectional
variation in capital structure decisions of international firms. Firms with more intense
informational asymmetries have a higher market leverage and they are more likely to issue
debt when raising new capital. Meanwhile, firms with less informational asymmetries are
more likely to issue equity and issue more equity. This occurs probably because the extent
of information asymmetry influences the comparative cost of their external capital. We
next demonstrate that, consistent with our expectation, the greater the adverse selection
problem, the higher the firms’ cost of equity. In contrast, firms’ cost of debt is not
significantly influenced by the adverse selection problem. This piece of evidence suggests
that shareholders require a higher premium than debtholders to cover the information cost
resulted from the adverse selection problem. Firms with more information asymmetries
attempt to avoid the large agency costs associated with equity capital and use more debt
capital instead. Our findings are robust to various sub-samples, as well as to alternative
model specifications and estimation techniques.

2
Bharath, Pasquariello, and Wu (2009) argue that, although broader liquidity measures capture com-
ponents besides adverse selection, such as inventory and order processing, the possibility that these
liquidity measures are derived from noninformational liquidity could be minimized through combining
them with other information asymmetry proxies based on their first principal component.
 3

Given the fact that firms with more intense information asymmetries have a higher
leverage ratio, we further examine what maturity structure is preferred by such firms.
Our results reveal that the extent of information asymmetry reduces firms’ use of long-
term debt. This piece of evidence is consistent with the findings of Barclay and Smith
(1995) on U.S. firms. The pricing of long-term debt is more sensitive to informational
asymmetries than short-term debt and thus long-term debt is associated with greater
information costs. As a result, firms with more information asymmetries are likely to use
more short-term debt while firms with less information asymmetries are likely to use more
long-term debt (Flannery (1986)).

Country institutions and infrastructure significantly impact the positive relation be-
tween information asymmetry and capital structure. We interact the effects of asymmetric
information with various institutional variables and find that the impact of information
asymmetry on leverage has been strengthened for firms in countries with developed bank-
ing systems as they may enhance the willingness of lenders to provide debt capital. In
contrast, the effect of information asymmetry has been mitigated for firms from coun-
tries with common-law origin and strict disclosure requirements that are documented to
facilitate equity issuance. Our results suggest that, in an international setting, institu-
tional arrangements influence the availability and cost of different types of capital and
then modify the extent to which information asymmetry could influence a firm’ capital
structure.

Our information asymmetry index is created from market-based liquidity measures.

However, Frieder and Martell (2006) document there exists a bi-directional relation be-
tween market liquidity and capital structure. To address the potential endogeneity prob-
lem, we employ a set of alternative information proxies, including an external information
event of IFRS(International Financial Reporting Standards) adoption, the characteristics
of analyst activities and earnings management, to check the robustness of our results.
We find that firms in countries that adopted IFRS experience significantly lower leverage
 4

following the IFRS mandatory enforcement, indicating an improvement in information

environment creates an incentive to use more equity than debt. Also, when analysts are
uncertain about firms’ future earnings or earnings opacity increases, i.e., the degree of
information asymmetry is high, firms generally have high market leverage and a lower
long-term debt ratio. In general, the results with alternative information proxies confirm
our previous findings.

Our study makes potential contributions to the existing literature. First, we provide
a comprehensive empirical test of the peking order theory and evidence suggests that in-
formation asymmetry creates a preference ranking over financing sources for international
firms. Furthermore, we find that the adverse effect on the cost of equity associated with
information asymmetry is not as much as the adverse effect on the cost of debt. It suggests
that information asymmetry influences a firm’s financing decisions due to its differential
impacts on the cost of different types of external capital. Second, we distinguish the ef-
fects of asymmetric information on capital structure by country characteristics and show
that firms tailor their financing choices to institutional features, which may strengthen or
weaken the effects of information asymmetry on capital structure depending on whether
a particular institutional feature facilitates debt or equity financing. Third, the potential
endogeneous relation between liquidity and capital structure has been carefully addressed.
Our basic results are robust against an external information event, i.e., the IFRS manda-
tory enforcement, and other widely-used dynamic proxies of adverse selection derived from
analyst forecasts and earnings management.

The paper is organized as follows. Section 2 reviews the related literature and develops
hypotheses. Section 3 describes the data and our sample construction. Sections 4 discusses
the relation between information asymmetry, capital structure, and the cost of capital and
Section 5 examines how this relation varies with the economic environment. Section 5
presents the robustness tests to alternative information proxies, and the final section
concludes the paper.
 5

2 Hypothesis Development

The pecking order theory suggests that informational asymmetries between firm manage-
ment and outside investors would create a financing hierarchy, i.e., firms would prefer
debt over equity when they need external financing (Myers and Majluf (1984)). Accord-
ing to the theory, equity capital -the most information-sensitive security- has the largest
adverse selection cost so firms prefer to raise equity as a financing means of last resort.
Consequently, firms with intense information asymmetry will have high leverage ratio.
In the past few decades, a large body of work has been devoted to scrutinize whether
pecking order accurately describes the observed financing behavior. One strand of studies
finds evidence in support of the pecking order theory and argue that it is a good approx-
imation of reality for some instances (See Shyam-Sunder and Myers (1999) and Fama
and French (2002)). The other strand of studies shows that the pecking order theory
fails under certain conditions (See Jung, Kim, and Stulz (1996), Frank and Goyal (2003),
Bolton and Dewatripont (2005), and Leary and Roberts (2010)). Recently, several studies
further attempt to evaluate the core assumption of pecking order theory, i.e., information
asymmetry as a determinant of capital structure decisions (See Bharath, Pasquariello,
and Wu (2009) and Agarwal and O’Hara (2007)). We investigate the role of information
asymmetry in explaining capital structure decisions for a sample of international firms
and conjecture that:

Hypothesis 1: Firms with a higher level of information asymmetry tend to have larger
market leverage.

In addition to the effect of information asymmetry on the overall capital structure

decisions, we are further interested in how a firm’s debt maturity structure is affected
by its information setting. In an environment characterized by informational asymmetry,
all the market’s required default premia are viewed to be excessive for relatively trans-
parent firms. However, the premium on long-term debt will appear most unreasonable
because the probability of credit quality deterioration assigned by the market is higher
 6

than expected by firm insiders(Flannery (1986)). Hence, firms with intense information
asymmetries will prefer short-term borrowing and firms with less information asymme-
tries will prefer long-term borrowing. Several empirical studies evaluate the relationship
between risk ratings, asymmetric information, and a firm’s debt maturity structure on
U.S. samples(see Barclay and Smith (1995), Stohs and Mauer (1996), Johnson (2003),
and Berger, Espinosa-Vega, Frame, and Miller (2005)). For example, Barclay and Smith
(1995) measure the information asymmetry by the amount of growth options and show
that firms with larger potential information asymmetries use more short-term debt. The
discussion leads to the second hypothesis:

H1a: Firms with a higher level of informational asymmetries tend to use less long-term
debt.

We next examine the effect of information asymmetry on the capital structure choices
made by firms when external funding is needed. Investigating incremental capital raising
decisions using discrete choice analysis allows us to assess the impact of information
asymmetry at points when firms decide to refinance or change their capital structure.
With asymmetric information, firms would prefer to issuing debt rather than equity when
investment needs exceed internal funds because the adverse selection costs associated with
the issuance of equity are larger than those with the issuance of debt. We identify financing
decisions by relative changes in debt and equity and calculate changes in amounts of
outstanding equity and debt for each firm in each year and expect that:

H1b: Firms with a higher level of informational asymmetries are more likely to issue
debt and tend to issue more debt when raising external capital.

H1c: Firms with a lower level of informational asymmetries are more likely to issue
equity and tend to issue more equity when raising external capital.

Concerning the cost of capital, Myers and Majluf (1984) state that equity capital is the
most information sensitive security and it is associated with the highest adverse selection
cost. By contrast, debt capital has much less adverse selection cost and internal funds
 7

completely avoid the problem. Consequently, the required adverse selection risk premium
by investors would be higher for equity than for debt. Other theoretical studies also
suggest an association between the information structure of a firm and the value of its
securities(see Barry and Brown (1985) and Coles, Loewenstein, and Suay (1995)). These
models predict that more information disclosure allows investors to better estimate asset
returns, which in turn lowers the required rate of return on assets. Empirical research has
provided strong evidence on the impact of firms’ information environment on the cost of
both equity and bond securities. For example, Easley and O’Hara (2004) and Lambert,
Leuz, and Verrechia (2007) find that information asymmetries adversely affect firms’ cost
of equity. Especially, an increase in the amount of information disclosed would lead to an
unambiguous decline in the cost of equity. Recently, it has been noted that information
concerning firms’ true value could be valued outside the equity market. Mansi, Maxwell,
and Miller (2011) report that information contained in analysts forecasts reduces bond
yield spreads, especially when the uncertainty about firm value is high. The discussion
leads to the following hypothesis:

Hypothesis 2: Both the cost of equity and the cost of debt increase with the ex-
tent of information asymmetries, but the cost of equity is more sensitive to information
asymmetries.

Our international setting is especially interesting as institutional and legal environ-

ments differ substantially across countries. It is highly possible that the relevant country-
level factors such as the development of financial sector, legal system origin, and disclo-
sure requirements may affect the association between information asymmetry and capital
structure. It is expected that:

Hypothesis 3: The effect of information asymmetry on capital structure vary with a

country’s institutional environment.

First, the development of banking systems can directly influence firms’ financial struc-
tures. Financial intermediaries have comparative advantage in collecting information and
 8

greater incentives to monitor firms than small investors (Fama (1985)). Thus, a sophis-
ticated banking system would facilitate firms, especially small firms, to access external
finance. We proxy the level of banking development by the amount of domestic credit
provided by banking sector (Demirgüc-Kunt and Levine (2001))and conjecture that:

H3a: The effect of information asymmetry on market leverage is larger in countries

characterized by a developed financial sector.

Second, the adoption of different legal systems has impact on corporate financial de-
cisions. Common-law countries tend to provide better protection to outside investors
against expropriation by insiders than civil law countries and thus higher valuations of
securities. Generally, common-law countries also have more developed and broader cap-
ital markets to which firms would easily access (Demirgüc-Kunt and Maksimovic (1998)
and La Porta et al. (1998)). As a result, it could be comparatively easier for firms to
raise external finance through equity in common-law countries than for firms in civil-law
countries. We define a legal origin dummy which equals one if a firm is from common law
oriented countries and zero if it is from countries with a civil-law origin. It is expected
that:

H3b: The effect of information asymmetry is smaller in common law countries.

Third, credible disclosures can mitigate information asymmetry between firm man-
agement and investors as well as among investors, which will lower firms’ cost of external
financing. The adverse selection problem may become less severe in countries with exten-
sive disclosure practice, where firms with great extent of information asymmetry could
be easily identified as ”bad” and lose the edge in markets when competing for capital.
We employ a widely-used disclosure index from the Center for International Financial
Analysis and Research database to represent both the mandated and voluntary disclosure
practices. Prior literature shows that countries with weak investor protection environ-
ment have lower quality auditing and accounting standards, thereby probably making the
disclosure unreliable and ineffective (see La Porta et al. (1998)). Thus, we expect the
 9

voluntary disclosure to be more common in a well protected market and this measure
reflects the overall transparency of the market. The discussion leads to:

H3c: The influence of information asymmetry is smaller in countries with stringent

disclosure requirement.

3 Data and Sample

Firms’ financial information is obtained from Worldscope while the stock trading infor-
mation for estimating information asymmetry measures is from Datastream and Taqtic
files. We follow the standard practice of excluding utility and financial firms because their
financing policies are affected by government regulations. We require each firm to have at
least two consecutive years’ financial data as our independent variables are lagged by one
year. For the period 1997 to 2007, our selection procedure results in a sample of 90,514
firm-year observations for 13,019 unique firms in 39 countries. The start date of 1997 is
chosen as the trading information becomes available from 1996 onwards.

3.1 Measures for Information Asymmetry

Empirically capturing the information environment around a firm, especially the informa-
tion asymmetry about a firm’s market valuation, is a difficult task in corporate finance.
Market microstructure theory suggests that better informed traders could affect the pro-
cess of price formation and discovery in a financial market and transmit the advanced
information into observed transaction data. The literature of market microstructure has
developed numerous ways to measure adverse selection based on the notion that the liq-
uidity of an asset is greatly affected by the intensity of information asymmetry underlying
the asset. In this study, we construct a composite index of information asymmetry for
each sample firm within each fiscal year between 1996 and 2007. The firm-level index is
calculated as the first principal component of four information asymmetry proxies, i.e., the
probability of informed trading (PIN), proportional quoted and effective bid-ask spreads,
 10

and Amihud illiquidity ratio. Bharath, Pasquariello, and Wu (2009) demonstrate that,
by extracting the common variation in these information asymmetry proxies, the possi-
bility that these measures are derived from factors other than adverse selection (such as
inventory costs, transaction costs, etc.) could be minimized.

The probability of information-based trading (PIN), is estimated using the approach by

Easley, Kiefer, and O’Hara (1996) and Easley, Hvidkjaer, and O’Hara (2010). It measures
the intensity of information asymmetry across equity investor groups and the greater the
estimated PIN, the greater is a firm’s extent of information asymmetry. The two measures
of bid-ask spreads, proportional quoted spread (Glosten (1987)) and proportional effective
spread (Chordia, Roll, and Subrahmanyam (2000)), are the most demonstrable estimators
to capture overall liquidity. Amihud’s illiquidity ratio (Amihud (2002)) is a volume-based
liquidity indicator and reflects the response of market price to order flow. Intuitively,
when a firm’s stock is liquid, a large trading volume is accompanied by a small price
change. Thus, a high illiquidity ratio indicates less liquidity due to the greater extent of
adverse selection around the firm.

We standardize each firm-level adverse selection proxy cross-sectionally by country-

year and the median values of these standardized measures across countries are reported
in Appendix B. All proxies of information asymmetry are positively correlated with each
other. To enhance the efficiency and efficacy of these measures in capturing information
asymmetry, we compute the first principal component, Inf Asy, of the correlation matrix
of these cross-sectionally standardized adverse selection estimators for each firm-year. On
average, the first principal component can account for as much as 64% of the corresponding
cross-sectional sample variance and the mean of the first eigenvalue is 2.55 which is much
greater than the second one, 0.80. We conclude that the first principal component serves as
a good estimator to represent the common variability in the four adverse selection proxies.
The index is positively correlated with each individual component. For example, the
index has a correlation of more than 90% with the two proportional bid-ask spreads and
 11

of more than 50% with the Amihud illiquidity measure and PIN. Also, each proxy enters
the index with a decent loading, which varies from 34% on PIN to 59% on proportional
quoted spread. So each information asymmetry proxy plays its role in the index. By
construction, the more severe the adverse selection problem is, the greater is the value
of the index. The median values of the information asymmetry index are presented in
the first column of Appendix C. It shows that there is much variation across countries.
Particularly, the median firm in China has the most intensive information asymmetry with
an index value of -0.172 while the one in Austria has the least information asymmetry
with a value of -0.706.

3.2 Other Variable Definitions

We measure firms’ capital structure by market leverage and debt maturity. Market lever-
age is defined as total debt divided by the market value of a firm. Total debt equals the
sum of short-term and long-term interest bearing debt. Market value is given by total
assets minus book equity plus market capitalization. Debt maturity choice is defined to
be the ratio of long-term debt to total debt (Fan, Titman and Twite (2010)).

The cost of debt capital is defined by the interest rate on a firm’s debt, i.e., the interest
expense for the year over the average short-term and long-term debt during the year. Our
cost of debt estimate reflects an average historic pre-tax interest rate based on each firm’s
debt financing pattern over time. We use this accounting based cost of debt measure
instead of the yield to maturity of the outstanding bonds because the trading data for
international corporate bonds are largely unavailable. This measure has some potential
caveats as it does not account for the country-specific tax structure and the possible
variation in reporting rules with respect to interest expense across countries. However,
Francis, Khurana, and Pereira (2005) show that, despite the possible measurement errors,
this proxy of cost of debt works well in capturing the impact of firms’ disclosure incentives
on their cost of capital.
 12

The cost of equity capital is derived by the modified price-earnings growth valuation
model developed by Easton (2004). Specifically, the implied cost of capital estimate for
each firm, rM P EG , is computed as the internal rate of return that equates the current
stock price to the present value of expected future sequence of abnormal earnings.

ˆ t+2 + rM P EG · dˆt+1 − eps

(eps ˆ t+1 )
Pt = 2
. (1)
rM P EG

where Pt is the firm’s stock price at date t, eps

ˆ t+1 and eps
ˆ t+2 are the one-year-ahead
and two-year-ahead earnings per share forecasts respectively, and dˆt+1 is the expected
dividends per share in period t + 1. The inputs of stock prices and forecasted earnings
per share are obtained from the I/B/E/S database and measured as of month +10 after
the fiscal year end.3 Dividend forecasts are obtained from datastream and computed as a
fraction of expected future earnings per share which equals the historic two-year average
of dividend payout ratios for each firm. As demonstrated by Botosan and Plumlee (2005)
and Hail and Leuz (2006), this measure is a reliable empirical proxy for cost of equity
capital and consistently associated with various firm-specific risk and country factors.

In accordance with the existing literature, we control for a set of firm-specific variables
that have been shown to affect capital structure and the cost of capital. Asset tangibility
is net property, plant & equipment over total assets. Size is measured by the natural
logarithm of total assets. Profitability is net income over total assets. Market-to-book
ratio, a proxy for growth opportunity, is defined as total assets minus book equity plus
market capitalization divided by total assets. Book leverage is computed as book debt
to total assets. Beta is defined as the covariance of the Morgan Stanley Capital Inter-
national (MSCI) country index monthly return with each stock’s monthly return divided
by the monthly return variance of the MSCI country index, and is estimated over the
past 60 months till June of each year. Forecasting bias is measured as the one-year-ahead
consensus earnings forecast minus the actual earnings reported in I/B/E/S.
3
We deliberately choose this time lag to ensure that analyst forecasts are publicly available and
incorporated into prices at the time we compute the cost of capital estimate. (See Hail and Leuz (2006)
for more details).
 13

Panel A of Appendix C presents the median values of firm-level variables by coun-

try. To prevent outliers from affecting our analysis, we winsorize all firm-level variables
at 1% and 99% of their own distributions. Appendix D provides the Pearson pairwise
correlation coefficients between firm characteristics. The results suggest that firm char-
acteristics could potentially influence financing choices. Specifically, tangibility and firm
size are positively related to market leverage while profitability and market-to-book ra-
tio are negatively correlated with leverage. More importantly, our proxy of information
asymmetry is positively associated with leverage ratio, which suggests that firms with
intensive information asymmetry use more debt.

4 Information Asymmetry, Capital Structure, and

Cost of Capital

In this section, we first evaluate the impacts of information asymmetry on firms’ financing
decisions and the cost of capital and then we examine how the institutional environment
impacts the relation between information asymmetry and capital structures, if there is
any. All the model specifications include industry and year fixed effects and the inferences
are adjusted for heteroscedasticity and corrected for clustered standard errors.

4.1 Information Asymmetry and Capital Structure

4.1.1 Information Asymmetry and Market Leverage

To examine how firms’ capital structures vary with the intensity of information asymme-
try, we begin by sorting sample firms into deciles according to our information asymmetry
index. The mean values of leverage for each decile, which are country-, industry-, and
year- adjusted, are presented in Table 1. Decile 1 represents the lowest level of information
asymmetry while decile 10 represents the highest level of information asymmetry. It shows
that the market leverage ratio is increasing monotonically with the extent of information
asymmetry. The adjusted market leverage ranges from -0.0248 (in decile 1) to 0.0355 (in
 14

decile 10). The difference between decile 10 and decile 1 is statistically significant with a
t-statistic of 25.45, suggesting a very strong relation between leverage and the extent of
information asymmetry.

We further examine the role of information asymmetry in determining a firm’ capital

structure by regressing market leverage on the information asymmetry index along with
firm and country-level variables. The regression estimates of different models are presented
in Table 2. In Model 1, we control for country-, industry-, and year- fixed effects and
show that the coefficient of the information asymmetry index is 0.023, significant at 1%
level. This suggests that firms prefer to use more debt when they are facing more severe
adverse selection problems, which is in support of Hypothesis 1. The coefficient estimates
of all other firm-level determinants of capital structures are statistically significant and
bear the expected sign. Specifically, firms with a larger proportion of tangible assets
have higher leverage as tangible assets could secure the claims of creditors in the event
of liquidation. Large firms use more debt probably because they have a less likelihood
of bankruptcy and lower financial distress costs. By contrast, profitability and market-
to-book ratio are inversely related to leverage. Profitable firms prefer to retain earnings
within the firm to avoid asymmetric information costs associated with external financing.
Firms with great growth potentials tend to avoid issuing debt, as they are subject to high
bankruptcy costs and their intangible assets result in no value in the event of bankruptcy
- the under-investment problem defined by Myers (1977).

Model 2 controls for country characteristics that have been documented to influence
firms’ financing patterns. The observed positive relation between information asymmetry
and leverage still remains highly significant. Regarding the various institutional variables,
firms in countries with more developed banking system and firms from common-law coun-
tries tend to use more debt. Also, firms exhibit a lower leverage ratio if they are from
countries with more transparent environment, where equity is a better option than debt.

We then investigate how robust our results are to different sample selections and
 15

research methods. First, we perform our test in a sample that excludes firm-year obser-
vations from Japan, as Japan accounts for the largest observations in our sample. Model
3 shows that Inf Asy still has a positive coefficient of 0.014 which is significant at 1%
level. Models 4 and 5 show that the positive association between information asymmetry
and leverage holds for the subs-periods of 1997 to 2001 and 2002 to 2007 as well. Due to
the constraint on stock trading information coverage, our sample extends from 1997 to
2007. We then replace our information asymmetry proxy with Amihud’s ratio which has
a better coverage back to 1990. Consistent with our expectation, the Amihud’s ratio is
positively related to a firm’s leverage, as shown in Model 6.

Up to this point our analysis has focused on the pooled time series, cross-sectional
regressions. However, the error terms could be correlated across firms as well as correlated
over years within a firm. To account for this potential correlation problem, we use the
approach proposed by Fama-MacBeth (1973) and present the average of coefficients and t-
statistics in Model 7. The information asymmetry index still has a positive and significant
effect on firm leverage. Its average coefficient is 0.017 and statistically significant at 1%.
Another concern is that the extent of information asymmetry could be influenced by
the financing patterns of a firm. Also, it is possible that information asymmetry and
capital structure are jointly determined. To address the endogeneity issue, we take the
difference in both dependent and independent variables and run regressions based on these
differences. Model 8 shows no material change to our previously documented results.
Finally, our dependent variable of leverage is limited between zero and one. We thus
employ a censoring regression model and the impact of information asymmetry on capital
structure becomes even more significant as shown in Model 9.

4.1.2 Information Asymmetry and Debt Maturity Choices

Univariate tests about the relation between information asymmetry and debt maturity
structure are displayed in Table 1. It appears that long-term debt is decreasing with the
extent of information asymmetry. The adjusted long-term debt ratio is 0.0826 in decile 1,
 16

compared to -0.0769 in decile 10. The difference of decile 10 minus decile 1 is significant
at 1% level, suggesting long-term debt is associated with a high information cost.

The regression results are reported in Table 3. In Model 1, the information asymmetry
index has a coefficient of -0.009 that is statistically significant at 1% level. Our finding
supports H1a and is in line with the argument of Flannery(1986) that firms with intense
adverse selection problems tend to use more short-term debt. The pricing of long-term
debt is more sensitive to changes in firm value than the pricing of short-term debt. Firms
with more information asymmetries are more concerned about the signalling effects of debt
maturity choice and likely to use less long-term debt because of the larger information
costs associated with it. With respect to control variables, it shows that firms with more
tangible assets in place and with large size use more long-term debt. Market-to-book also
presents a positive relationship with long-term debt.

Model 2 examines the impact of country characteristics on firms’ debt maturity choices.
It shows that firms from common law countries or countries with extensive disclosure prac-
tice tend to use more long-term debt. This could be explained that these two institutional
features could enhance the transparency of information environment surrounding a firm
which reduces the cost associated with the mispricing of long-term debt. The results
with sensitivity checks do not differ significantly from those observed in Models 1 and 2.
Our baseline evidence that information asymmetries are negatively associated with firms’
debt maturity structure is robust against various subsamples, subperiods, and alternative
estimation techniques.

4.1.3 Information Asymmetry and Capital Structure Choices

This subsection investigates the impact of information asymmetries on the financing choice
variables. We identify capital structure choices by computing the relative changes in
amounts of outstanding equity and debt for each firm in each year. Amount of equity
issued is defined as sale of common and preferred stock minus any purchase of common
 17

and preferred stock, scaled by total assets. Amount of debt issued is calculated as long-
term debt issuance minus long-term debt retirement, scaled by total assets. Equity issues
dummy equals 1 if a firm’s net equity issues are equal to or greater than 5% in a year and
0 if otherwise, while debt issues dummy equals 1 if a firm’s net long-term debt issues are
greater than 0 in a year, and 0 if otherwise.

The results are presented in table 4. It shows that information asymmetry has a signif-
icant impact on firms’ financing choices when they raise external capital. Particularly, the
information asymmetry index has negative coefficients of -0.005 (-0.173) in regressions on
equity issuance (equity issuance dummy). This suggests that firms with intense informa-
tion asymmetry have less equity issuance and are less likely to issue equity. However, the
information asymmetry index has significantly positive coefficient in regressions on debt
issuance and debt issuance dummy. Alternative models with country fixed effects and
country characteristics are employed as well and the results remain consistent. Overall,
the evidence is consistent with the predictions of H1b and H1c, indicating that the extent
of information asymmetry will lower and enhance firms’ propensity to issue equity and
debt, respectively.

4.2 Information Asymmetry and Cost of Capital

The previous sections have documented that information asymmetry exhibits a substan-
tial impact on firms’ capital structure. We next examine the issue of how information
asymmetries impact the cost of capital. The univariate tests of Table 1 report that the
cost of equity is strictly increasing with the level of information asymmetry. The differ-
ence of the adjusted cost of equity between decile 10 (highest information asymmetry) and
decile 1 (lowest information asymmetry) is 0.0807, with a t-statistic of 20.52. The cost
of debt, on the contrary, appears to be not very responsive to the changes of information
asymmetry. The difference of decile 10 minus decile 1 is not statistically significant.

The regression results presented in Table 5 substantiate the findings from univariate
 18

analysis. Models 1 and 3 indicate that information asymmetry is positively associated

with the cost of debt but without statistical significance. Considering control variables,
profitability is negatively associated with the cost of debt. The likelihood of default for
profitable firms is lower. Consequently, the associated cost of debt would be lowered as
well. Book leverage has a negative impact on the cost of debt. Probably, firms prefer to
use more debt when the cost of debt is relatively low and thus leverage ratio represents a
cumulative effect of the low cost of debt rather than a proxy for default risk.4 To capture
the difference in the base interest rates across countries, we follow Francis, Khurana, and
Pereira (2005) and include a control variable of country-specific bank rate in the regression
models. It denotes the bank rate that applies to the short-and medium-term financing
in the private sector and obtained from the International Financial Statistics. Results
show that the lending rate has no significant effect on the cost of debt in our sample.
More country-specific variables are included in Model 3. However, none of them exhibit
significant influence on the cost of debt.

Models 4 and 6 display the regression estimates for the cost of equity. It shows
that more information asymmetry surrounding a firm is associated with higher equity
cost. Traditional risk proxies, such as size, market-to-book, beta, and leverage carry the
expected signs and have a high level of statistical significance. Inflation rate, computed
as the median of the current year’s annualized monthly inflation rates, is included as the
forecast earnings are expressed in nominal terms. It has a significant coefficient of 0.001
that falls into the range of zero to one as suggested by Hail and Leuz (2006). As for the
institutional factors, the development of banking sector and the disclosure practice are
negatively related to the cost of equity.

Finally, we employ two approaches to facilitate the direct comparison between the pos-
sible impacts of information asymmetry on the cost of debt and the cost of equity.We first

4
However, we find that the cost of debt is positively associated with the square term of book leverage,
indicating a nonlinear relationship exists between the cost of debt and leverage ratio and a high leverage
ratio still implies a high level of default risk.
 19

use standardized regressions as reported in Models 2 and 5. The standardized coefficient

of information asymmetry is 0.002 for the cost of debt and 0.085 for the cost of equity.
As discussed earlier, the higher the degree of informational asymmetries, the higher the
cost of equity and the cost of debt. However, the impact of information asymmetry on
the cost of equity is much stronger than its impact on the cost of debt. Second, we create
a new variable of the standardized cost of debt minus the standardized cost of equity. As
expected, Models 7 and 8 show that information asymmetry has a significant and negative
effect on the cost of capital differential, suggesting that the cost of equity is more sensitive
to a firm’s information environment.

Overall, the results of table 5 provide strong support for Hypothesis 2. Understanding
how the impacts of information asymmetry vary across different types of capital, firms
exhibit the pecking order behavior when they have a severe adverse selection problem. We
thus conclude that the extent of information asymmetry plays a crucial role in determining
firms’ capital structure by affecting the comparative costs of their external financing.

4.3 When Does Information Asymmetry Matter Most?

Evidence from previous sections implies that the degree of information asymmetry in-
fluences firms’ financing decisions through its comparative impacts on different types of
capitals, as stated in the pecking order theory. However, in the international setting, in-
stitutional features and country-level information environment differ substantially across
countries. In this section, we attempt to explore whether and how various institutional
variables could affect the impacts of information asymmetry on firms’ financing choices.

In Table 6, we regress the market leverage on the information asymmetry index to-
gether with the product of information asymmetry and institutional variables. The in-
teraction terms could capture any influences of country factors on the observed effects of
information asymmetry on capital structures. Model 1 shows that the interaction term
between information asymmetry and the strength of banking sector has a significant co-
 20

efficient of 0.004, indicating that the impact of information asymmetry varies with the
development of financial sector. The information asymmetry surrounding a firm has a
larger effect on leverage ratio as raising equity might be more difficult than raising debt
in countries with more developed banking sector.

Model 2 shows that the effect of information asymmetry has been alleviated for firms
located in common-law countries rather than civil-law countries. The interaction term
between information asymmetry and the legal origin dummy has a significantly negative
coefficient of -0.004. Common-law countries are perceived to have good shareholder pro-
tections and more developed capital markets, both of which facilitate equity financing.
Model 3 shows that the coefficient of the interaction of information asymmetry and the
disclosure proxy is -0.023, suggesting firms are inclined to use less debt if they are located
in countries with rigorous disclosure requirements. With a transparent external environ-
ment, shareholders can obtain more information about firms’ true value, which makes the
equity issuance less costly. Finally, we include all the interaction terms between informa-
tion asymmetry and country characteristics in Model 4 and the results remain materially
unchanged.

Generally, when a country is characterized to be more favorable to equity financing,

firms would take this opportunity and use more equity than what they might use oth-
erwise. This is in support of H1a, H1b and H1c. Specifically, equity issuance of firms
from countries that have common law origin and stringent disclosure requirements are less
sensitive to the surrounding information environment as they are less concerned about
the agency costs associated with information asymmetry. However, it is worth noting
that the pecking order still exists as the coefficient of information asymmetry index is
positive and significant at 1% level. When a country is characterized to be less favorable
to equity financing, such as more developed banking sector, firms’ equity issuance will be
more affected by the agency costs associated with the information asymmetries between
firm insiders and outside shareholders.
 21

5 Alternative Information Proxies

Our information asymmetry measures are market-based, calculated from liquidity mea-
sures or from information contained in stock transactions. However, Frieder and Martell
(2006) put forth an idea that the relation between liquidity and leverage could move bi-
directionally. Increases in debt will increase firms’ financial burden (for example, interest
and principal payments) and risk level, which may cause firm managers, with minimal
equity share, to avoid risky investment projects and not make value-maximizing deci-
sions. The stock will be less attractive to investors, resulting in wider bid-ask spreads
or reduced liquidity. To address this concern, we now examine the robustness of our re-
sults to alternative information proxies. The first one is an external information event,
i.e.,the mandatory enforcement of International Financial Reporting Standards (IFRS)
in European Union countries in 2005. The second set of information proxies is created
from the information contained in analyst activities, followed by information asymmetry
concerning earnings opacity.

5.1 Mandatory Enforcement of IFRS

Starting on 1 January 2005, the European Union (with certain exceptions) requires pub-
licly traded companies to present consolidated financial statements in conformity with
IFRS, a set of high quality accounting standards.5 By then, 14 EU member countries
including Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy,
Netherlands, Portugal, Spain, Sweden and UK have adopted IFRS. At the same time,
IFRS or certain accounting regime equivalent to IFRS is also legally required by some
countries with prominent capital markets, such as Australia, Hong Kong, Singapore and
South Africa. The adoption of IFRS enhances the credible implementation of uniform
5
Firms within the EU that are entitled to follow alternative accounting standards include companies
listed in the Alternative Investment Market (AIM) and Swiss firms that are not multinationals. However,
the AIM has adopted a rule that requires AIM firms to submit IFRS financial statements for periods
beginning on or after 1 January 2007, although voluntary adoption is allowed (Horton, Serafeim, and
Serafeim (2008)).
 22

accounting standards and has the potential to facilitate cross-border financial statement
comparability. A direct consequence would be a decrease in information acquisition costs,
thereby increasing competition and efficiency in the markets (Ball (2006)). Horton, Ser-
afeim, and Serafeim (2008) document that the transparency of information environment,
proxied by analyst activities, has been improved on average for non-financial firms dur-
ing the period around the mandatory adoption of IFRS. Moreover, this is true not just
for mandatory adopters but also for firms that voluntarily adopt IFRS earlier and even
non-adopters that continue to report under other GAAP after 2005. Therefore, this ex-
ogeneous event provides an ideal setting for assessing the possible impact of information
asymmetry on capital structure.

To examine the possible impacts of the mandatory enforcement of IFRS on capital

structures and the cost of capital, we focus on the subsample of firms that are incorporated
in one of the 19 countries mentioned above and employ two different approaches. The
results are presented in Table 7. In panel A, we define a dummy variable, IF RS, that
equals one for year 2007 and 0 before 2005 and then construct an interaction term between
the information asymmetry proxy and IFRS.6 Some interesting results emerge. First, the
IFRS enforcement has material effects on firms’ financing decisions. Market leverage
and long-term debt ratio exhibit significant decrease and increase, respectively, after the
IFRS adoption as in Models 1 and 2. Specifically, market leverage is lowered by 0.004
while long-term debt ratio is increased by 0.014. The coefficient on the interaction term
is marginally significant for market leverage, with a magnitude of -0.003, but it is not
significant for the long-term debt ratio. This suggests that, with a more transparent
information environment, the agency costs associated with information asymmetry has
been reduced and thus the market leverage has been reduced accordingly. Secondly,
the effect of information asymmetry on both the cost of debt and the cost of equity
has decreased after the IFRS enforcement, but only the decrease for the cost of equity

6
Years 2005 and 2006 are excluded as the effects of IFRS may not be observed immediately after the
adoption.
 23

is significant. Models 4 and 5 show that the interaction term between IFRS and the
information asymmetry proxy has a nonsignificant coefficient of -0.037 for the cost of
debt and significant coefficient of -0.010 for the cost of equity, respectively. This finding
implies a significant reduction in the cost of equity after the mandatory enforcement of
IFRS.

In panel B, the dependent and independent variables are measured as the difference
over the time window of 2004 to 2007, that is, year 2007 minus year 2004. The results
are in line with our predictions. Particularly, a reduction in information asymmetry is
followed by a reduction in market leverage and an increase in long-term debt ratio. As
for the cost of capital, a reduction in information asymmetry results in a reduction both
in the cost of debt and in the cost of equity. This indicates that the adoption of IFRS
possibly reduces the adverse selection risk for market participants as firms disclose more
information by conforming to uniform IFRS standards, leading to a more efficient firm
valuation.

Overall, we reach a conclusion that the mandatory enforcement of IFRS is associated

with a decrease in market leverage as a result of a decrease in cost of equity capital.
More specifically, the adoption of IFRS standards results in greater conformity in terms
of financial reporting which can mitigate information asymmetry surrounding a firm. As
a result, firms would issue more equity when they need external finance because the
information-sensitive capital becomes less expensive.

5.2 Analyst Activities and Earnings Management

Our information asymmetry measures are market-based, calculated from liquidity mea-
sures or from information contained in stock transactions. However, Frieder and Martell
(2006) put forth an idea that the relation between liquidity and leverage could move bi-
directionally. Increases in debt will increase firms’ financial burden (for example, interest
and principal payments) and risk level, which may cause firm managers, with minimal
 24

equity share, to avoid risky investment projects and not make value-maximizing decisions.
The stock will be less attractive to investors, resulting in wider bid-ask spreads or reduced
liquidity. To address this concern, we now examine the robustness of our results to alter-
native information proxies created from the information contained in analyst activities,
followed by information asymmetry concerning earnings opacity.

As an integral part of the capital market, financial analysts provide earnings forecasts,
buy/sell recommendations and other information to brokers, money managers and insti-
tutional investors. We include the number of financial analysts following a firm and two
properties of the analysts’ earnings forecasts (see Lang and Lundholm (1996)). One is the
degree of dispersion among forecasts which is defined as the standard deviation of analyst
forecasts scaled by the mean dispersion of analyst forecasts, and the other is analyst fore-
cast errors which are measured by the absolute value of the difference between announced
earnings and the average estimated earnings scaled by the average analyst forecasts. Ex-
tensive analyst coverage improves the transparency of a firm by allowing firm-specific
information to disseminate in a more timely manner. Greater forecasting dispersion and
higher forecasting errors may suggest low transparency or high uncertainty about a firm’s
future earnings and cash flows. The results are reported in Table 8. As expected, we find
that firms with extensive analyst coverage use less debt in aggregate but more long-term
debt (Models 1 and 7). In addition, firms with greater analyst dispersion and higher
forecast errors exhibit higher market leverage (Models 2 and 3) and lower long-term debt
ratio (Models 8 and 9).

Our analysis also employs three measures to characterize the information contained in
earnings management: the magnitude of accruals divided by operation cash flow,7 earn-
ings smoothing - the ratio of the standard deviation of operating income to the standard

7
Accruals = (∆CA - ∆Cash) - (∆CL - ∆SD - ∆TP) - Depr, where ∆CA is a change in total current
assets; ∆Cash is a change in cash and equivalent; ∆CL is a change in total current liability; ∆SD is a
change in short-term debt included in current liabilities; ∆TP is a change in income taxes payable; Depr
is depreciation and amortization expenses; when short-term debt and taxes payable are not available for
a firm, their changes are assumed to be zero; All accounting variables are scaled by lagged total assets
and measured in year t-1.
 25

deviation of cash flows over the last 5 years, and earnings correlation - the correlation
coefficient between the change of accruals and the change of operating cash flows over the
last 5 years (Leuz, Nanda, and Wysocki (2003)). Earnings accruals measure the extent to
which insiders exercise discretion in the accounting component of reported earnings. Earn-
ings smoothing measures the extent to which insiders reduce the variability of reported
earnings by altering the accruals. Earnings correlation captures the extent to which man-
agers use their accounting discretion to conceal economic shocks to their firm’s cash flows
from operations. Greater earnings volatility leads to a greater informational advantage
for informed investors over uninformed investors and, smoothed reported earnings would
reduce the trading loss for liquidity investors (Goel and Thakor (2003)). We expect a
positive relation between information asymmetry and accruals, but a negative relation
between information asymmetry and both earnings smoothing and earnings correlation.
Models 4 and 10 of Table 6 show that firms with a large amount of earnings accruals have
a higher market leverage ratio but a lower long-term debt ratio. However, the association
between accruals and the long-term debt ratio is not statistically significant. Earnings
smoothing and earnings correlation, on the other hand, lead to lower market leverage
(Models 5 and 6)and higher long-term debt ratio (Models 10 and 11). With alternative
information proxies, we still obtain the similar results as documented before.

6 Conclusion

We examine the interdependencies of information asymmetry, capital structure and the

cost of capital for a sample of firms from 39 developing and developed countries world-
wide. Information asymmetry positively affects firms’ leverage, possibly because infor-
mation asymmetry adversely affects the cost at which firms could access to equity but
no significant effect on the cost of debt capital has been found. As a result, firms at-
tempt to use more debt to minimize their overall cost of capital. In particular, firms with
intensive information asymmetry tend to use more short-term debt because the cost of
 26

long-term debt is more sensitive to the information environment. More interestingly, the
importance of information asymmetry in determining the observed corporate financing be-
havior is influenced by various country-specific factors. The positive association between
information asymmetry and market leverage is more pronounced for firms in countries
with strong banking sectors. On the other hand, the effect of information asymmetry has
been reduced if firms are located in common-law countries or countries with strong disclo-
sure practice. In sum, our evidence suggests that information asymmetry is an important
determinant of firms’ financing behavior in the international context and the importance
of information asymmetry is influenced by institutional environment.
 27

Appendix A. Variable Definition

VARIABLES ACRONYM DESCRIPTION

(i) Firm Characteristics

Information asymmetry Inf Asy The first principal component calculated from the adverse selection measures of
probability of informed trading (PIN), proportional quoted spread, proportional effective spread
and Amihud illiquidity ratio.
Market leverage DebtM Total debt over the sum of total debt and total market value of equity.
Debt maturity LT D Long-term debt over total debt.
Cost of debt COD Interest expense scaled by the average of short-term and long-term debt during the year.
Cost of equity COE The implied cost of equity capital from Easton (2004).
Tangibility FA Net fixed assets over total assets.
Size Size Logarithm of total assets in US$ millions.
Return on assets ROA Net income over total assets.
Market-to-book MB Total assets minus book equity plus market capitalization over total assets.
Book leverage DebtB Total debt over total book assets.
Beta Beta The covariance of Morgan Stanley Capital International(MSCI) country index monthly return
with each stock’s monthly return divided by MSCI country index monthly return variance over
the past 60 months till June of each year t.
Forecasting bias F Bias One-year-ahead consensus forecast minus the actual earnings reported in I/B/E/S.

(ii) Country Characteristics

Banking sector Bank Domestic credit provided by banking sector scaled by GDP.
Legal origin Law Indicator variable which equals 1 if a country is from the common-law origin
and 0 if it is from the civil-law origin.
disclosure requirements index, public enforcement index, and liability standard index.
Accounting Transparency CIF AR Measures the quality of accounting disclosure practice obtained from the Center
for International Financial Analysis and Research data base.
Inflation Inf l Median of the current year’s annualized monthly inflation rate.
Lending rate rlend Bank rate that meets the short-and medium-term financing needs of the private sector
obtained from International Financial Statistics.

(iii) Analyst Variables

Analyst coverage # of Analyst The number of financial analysts covering a firm

Forecast dispersion F disp Standard deviation of analyst forecasts scaled by the mean dispersion of analyst forecasts.
Forecast errors F Err Absolute value of the difference between announced earnings and the mean of
estimated earnings scaled by the mean of analyst forecasts.

(iv) Earnings Management

Accruals Accrual Magnitude of accruals divided by operating cash flows.
Earnings smoothing ESmooth A ratio of the standard deviation of operating income to the standard deviation
of cash flows over the last 5 years.
Earnings correlation Ecorr Correlation coefficient between the change of accruals and the change of operating cash flows
over the last 5 years.
 28

Appendix B
Components of the Information Asymmetry Index
This table presents the median values of the four components of information asymmetry index by country. P IN is
probability of informed trading from Easley, Kiefer, and O’Hara (1996). P RSprd is the proportional quoted spread
from Glosten (1987). P RESprd is the effective spread as in Chordia, Roll, and Subrahmanyam (2000). Amihud is
the Amihud illiquidity ratio from Amihud (2002). N obs is number of observations. Total indicates the median values
of the adverse selection and liquidity measures for the whole sample.

Country PIN PRSprd PESprd Amihud Nobs

Argentina 0.393 0.028 0.026 4.814 344
Australia 0.285 0.025 0.027 5.338 4891
Austria 0.219 0.014 0.015 0.670 333
Belgium 0.265 0.013 0.016 1.418 458
Brazil 0.313 0.052 0.047 10.211 151
Canada 0.241 0.020 0.020 1.101 2836
Chile 0.364 0.038 0.031 5.170 424
China 0.172 0.002 0.003 0.032 5250
Denmark 0.298 0.018 0.017 0.814 620
Finland 0.256 0.012 0.012 0.660 601
France 0.238 0.013 0.020 1.862 3321
Germany 0.177 0.024 0.016 2.378 2773
Greece 0.211 0.011 0.014 0.476 1505
Hong Kong 0.308 0.020 0.022 2.274 4511
India 0.223 0.006 0.006 0.914 2719
Indonesia 0.378 0.037 0.038 39.799 1839
Ireland 0.243 0.024 0.020 0.727 182
Israel 0.211 0.011 0.012 0.280 462
Italy 0.212 0.006 0.008 0.093 1047
Japan 0.214 0.008 0.008 0.145 18900
Korea 0.214 0.007 0.009 0.109 2189
Malaysia 0.295 0.011 0.011 1.866 3155
Mexico 0.325 0.022 0.021 3.829 444
Netherlands 0.208 0.009 0.010 0.123 642
New Zealand 0.361 0.012 0.014 1.128 294
Norway 0.281 0.016 0.016 0.837 602
Philippines 0.331 0.041 0.042 37.890 729
Poland 0.246 0.013 0.015 4.043 225
Portugal 0.275 0.012 0.014 3.303 168
Singapore 0.288 0.020 0.020 2.339 2125
South Africa 0.292 0.019 0.023 6.584 1221
Spain 0.195 0.006 0.007 0.047 589
Sweden 0.233 0.011 0.011 0.343 1275
Switzerland 0.278 0.010 0.015 0.569 595
Taiwan 0.208 0.005 0.006 0.038 4144
Thailand 0.311 0.016 0.017 10.138 2424
Turkey 0.196 0.013 0.013 0.153 1165
U.K. 0.224 0.034 0.020 0.550 5896
U.S. 0.147 0.004 0.003 0.005 9465
Total 0.227 0.010 0.010 0.270 90514
 Appendix C
Key Firm and Country Characteristics
This table shows the median values of the key firm- and country-level variables. Inf Asy is the information asymmetry index. DebtM is market leverage. LT D
is long-term debt ratio. COD is the cost of debt capital. COE is the implied cost of equity capital. F A is tangibility ratio. Size is the logarithm of total
assets. ROA is the return on total assets. M B is the market to book ratio. DebtB is book leverage. Beta is market beta from June of year t. F Bias is the
forecasting bias. Bank is the domestic credit provided by banking sector over GDP. Law is the legal origin dummy. CIF AR measures the level of accounting
disclosure practice. Inf l is the inflation rate in percentage. rlend is the lending rate. N obs is number of observations. Total is the median values of firm- and
country-level characteristics for the whole sample. All variables are detailed in Appendix A and firm-level variables are winsorized at 1% and 99% level. The
sample period is from 1997 to 2007.

Panel A: Firm Characteristics Panel B: Country Characteristics

Country Inf Asy DebtM COD COE LDT FA Size ROA MB DebtB Beta F bias Bank Law CIF AR Inf l rLend N obs

Argentina -0.26 0.18 0.14 0.20 0.47 0.49 11.47 0.03 0.91 0.20 0.45 0.05 0.37 0 0.76 8.83 0.11 344
Australia -0.34 0.03 0.08 0.13 0.69 0.29 9.69 -0.03 1.53 0.07 0.94 0.01 1.09 1 0.88 2.67 0.09 4891
Austria -0.71 0.20 0.06 0.13 0.54 0.36 12.25 0.04 1.21 0.25 0.68 0.05 1.23 0 0.66 2.06 0.06 333
Belgium -0.54 0.18 0.06 0.12 0.55 0.27 12.03 0.03 1.27 0.24 0.61 0.06 1.12 0 0.75 1.82 0.07 458
Brazil -0.44 0.15 0.21 0.16 0.48 0.29 12.82 0.05 1.28 0.26 0.70 27.35 0.78 0 0.65 6.84 0.55 151
Canada -0.46 0.07 0.07 0.15 0.72 0.40 11.13 0.01 1.59 0.12 0.88 0.02 2.09 1 0.81 2.14 0.06 2836
Chile -0.50 0.16 0.06 0.12 0.66 0.49 12.15 0.05 1.18 0.21 0.62 0.72 1.01 0 0.81 3.05 0.07 424
China -0.17 0.16 0.05 0.10 0.06 0.34 12.04 0.02 1.59 0.29 0.20 0.01 1.37 0 1.82 0.06 5250
Denmark -0.66 0.16 0.06 0.13 0.61 0.31 11.58 0.04 1.23 0.25 0.45 0.02 1.61 0 0.78 1.89 0.08 620
Finland -0.39 0.14 0.05 0.14 0.68 0.25 11.81 0.05 1.36 0.22 0.24 0.00 0.69 0 0.87 1.56 0.05 601
France -0.46 0.16 0.05 0.12 0.57 0.16 12.13 0.04 1.29 0.21 0.62 0.00 1.05 0 0.84 1.69 0.07 3321
Germany -0.36 0.13 0.08 0.16 0.55 0.19 11.52 0.02 1.22 0.17 0.54 0.04 1.38 0 0.73 1.58 0.10 2773
Greece -0.49 0.21 0.06 0.11 0.31 0.33 11.45 0.03 1.23 0.27 0.30 0.03 0.85 0 0.66 3.37 0.09 1505
Hong Kong -0.45 0.15 0.06 0.16 0.29 0.29 11.53 0.03 1.04 0.16 0.75 0.02 1.43 1 0.80 -0.40 0.07 4511
India -0.40 0.22 0.09 0.14 0.66 0.40 11.96 0.06 1.26 0.30 0.87 0.20 0.57 1 0.60 4.39 0.12 2719
Indonesia -0.48 0.34 0.09 0.19 0.42 0.41 11.21 0.03 1.05 0.35 0.74 14.08 0.50 0 10.45 0.17 1839
Ireland -0.66 0.14 0.06 0.11 0.70 0.18 12.57 0.04 1.54 0.22 0.53 0.00 1.36 1 0.88 3.94 0.03 182
Israel -0.53 0.20 0.06 0.12 0.61 0.24 12.55 0.03 1.24 0.26 0.90 0.01 0.81 1 0.76 1.12 0.07 462
Italy -0.31 0.19 0.05 0.12 0.51 0.19 12.90 0.02 1.25 0.25 0.81 0.01 1.02 0 0.73 2.09 0.06 1047
Japan -0.53 0.21 0.02 0.10 0.41 0.30 13.04 0.01 1.04 0.22 0.96 2.15 4.19 0 0.78 -0.25 0.02 18900
Korea -0.33 0.25 0.06 0.16 0.31 0.37 12.19 0.03 0.88 0.24 0.72 -24.00 1.07 0 0.75 2.54 0.06 2189
Malaysia -0.41 0.20 0.06 0.14 0.36 0.40 11.60 0.04 1.02 0.21 0.96 0.00 1.50 1 0.85 2.03 0.06 3155
Mexico -0.51 0.16 0.10 0.14 0.71 0.53 13.29 0.04 1.01 0.17 0.61 0.06 0.50 0 0.76 4.55 0.08 444
Netherlands -0.53 0.14 0.07 0.13 0.69 0.20 12.84 0.05 1.47 0.22 0.73 -0.01 1.70 0 0.79 1.70 0.04 642
New Zealand -0.58 0.15 0.08 0.12 0.82 0.41 11.41 0.06 1.47 0.24 0.76 0.00 1.17 1 0.86 1.82 0.10 294
Norway -0.51 0.19 0.07 0.18 0.81 0.23 11.97 0.03 1.31 0.26 0.88 0.10 0.90 0 0.84 2.26 0.07 602
Philippines -0.49 0.19 0.09 0.16 0.45 0.44 11.27 0.01 0.95 0.20 0.81 0.06 0.60 0 0.71 5.95 0.10 729
Poland -0.51 0.10 0.07 0.13 0.36 0.34 10.74 0.05 1.46 0.17 0.64 0.06 0.40 2.11 0.06 225
Portugal -0.69 0.30 0.05 0.11 0.54 0.35 12.72 0.02 1.09 0.36 0.57 0.01 1.43 0 0.60 2.74 0.07 168
Singapore -0.50 0.16 0.05 0.14 0.37 0.33 11.41 0.04 1.07 0.19 1.14 0.00 0.85 1 0.85 1.02 0.05 2125
South Africa -0.38 0.10 0.13 0.16 0.54 0.24 11.54 0.07 1.24 0.14 0.79 0.03 1.93 1 0.83 5.34 0.14 1221
Spain -0.51 0.18 0.07 0.11 0.53 0.33 12.95 0.04 1.30 0.25 0.68 0.02 1.32 0 0.76 3.04 0.05 589
Sweden -0.37 0.11 0.06 0.14 0.70 0.17 11.60 0.05 1.53 0.17 0.69 0.06 1.12 0 0.91 1.36 0.05 1275
Switzerland -0.51 0.14 0.05 0.12 0.64 0.31 12.41 0.04 1.23 0.18 0.77 -0.12 1.78 0 0.83 0.80 0.03 595
Taiwan -0.28 0.21 0.04 0.13 0.34 0.35 12.26 0.04 1.12 0.24 0.89 0.14 1.34 0 0.69 0.01 0.02 4144
Thailand -0.65 0.28 0.06 0.16 0.30 0.42 10.98 0.05 1.00 0.30 0.49 0.05 1.23 1 0.69 2.24 0.07 2424
Turkey -0.25 0.13 0.19 0.22 0.21 0.36 11.42 0.05 1.27 0.18 0.86 0.01 0.56 0 0.65 25.30 1165
United Kingdom -0.43 0.08 0.07 0.13 0.62 0.20 11.54 0.03 1.54 0.13 0.79 -0.05 1.57 1 0.91 2.93 0.05 5896
United States -0.52 0.14 0.07 0.10 0.92 0.26 6.99 0.05 1.55 0.24 0.79 -0.01 2.49 1 0.82 2.68 0.07 9465

Total -0.41 0.16 0.05 0.12 0.49 0.30 11.82 0.03 1.22 0.21 0.79 0.01 1.12 0 0.78 2.14 0.07 90514
29
 Appendix D
Pearson Pairwise Correlation Coefficients
This table presents the Pearson correlation coefficients for the key firm- and country-level variables. Inf Asy is the information asymmetry index. DebtM is
market leverage. LT D is long-term debt ratio. COD is the cost of debt capital. COE is the implied cost of equity capital. F A is tangibility ratio. Size is
the logarithm of total assets. ROA is the return on total assets. M B is the market to book ratio. DebtB is book leverage. Beta is market beta from June of
year t. F Bias is the forecasting bias. Bank is the domestic credit provided by banking sector over GDP. Law is the legal origin dummy. CIF AR measures
the level of accounting disclosure practice. Inf l is the inflation rate in percentage. rlend is the lending rate. All variables are detailed in Appendix A and
firm-level variables are winsorized at 1% and 99% level. The sample period is from 1997 to 2007.

DebtM COD COE LT D FA Size ROA MB DebtB Beta F bias Bank Law CIF AR Inf l(%) rLend

DebtM 0.084
COD -0.001 -0.015
COE 0.174 -0.088 0.007
LT D -0.145 -0.022 0.234 0.091
FA 0.000 0.170 0.000 0.271 -0.004
Size -0.371 -0.074 -0.003 0.193 -0.078 0.103
ROA -0.190 0.050 0.003 -0.011 -0.311 0.050 0.199
MB -0.022 0.036 0.009 -0.219 -0.088 -0.108 -0.174 -0.232
DebtB 0.042 0.126 -0.019 0.730 0.147 0.191 0.104 -0.114 -0.033
Beta -0.015 0.011 0.001 -0.006 0.058 -0.017 -0.002 -0.027 0.005 -0.005
F bias 0.015 -0.019 0.026 0.031 0.027 0.001 -0.002 -0.017 0.001 0.043 0.000
Bank 0.003 0.011 -0.021 0.041 -0.172 -0.058 0.083 0.028 -0.075 0.002 0.006 0.003
Law -0.003 0.229 -0.007 -0.115 0.018 0.049 -0.447 -0.082 0.098 -0.072 -0.017 -0.011 0.473
CIF AR 0.003 0.137 -0.022 -0.191 -0.044 -0.101 -0.171 -0.128 0.086 -0.113 -0.018 0.006 0.313 0.312
Inf l(%) 0.000 -0.009 0.028 0.023 0.167 0.040 -0.084 0.018 0.028 0.042 -0.006 0.026 -0.168 0.004 -0.181
rLend -0.002 0.067 0.088 0.040 0.163 0.075 -0.183 -0.024 0.080 0.054 0.002 0.013 -0.013 0.182 -0.117 0.651
30
 31

References

Agarwal, Prasun, and Maureen O’Hara, 2007, Information Asymmetry and Capital
Structure, Working Paper at Cornell University.

Amihud, Yakov, 2002, Illiquidity and Stock Returns: Cross-section and Time-series
Effects, Journal of Financial Markets 5, 31-56.

Ball, Ray, 2006, IFRS: Pros and Cons for Investors., Accounting and Business Research,
International Accounting Policy Forum, 5-7.

Barclay, Michael J., and Clifford W. Smith, 1995, The Maturity Structure of Corporate
Debt, Journal of Finance 50, 609-631.

Barry, Christopher, and Stephen Brown, 1985, Differential Information and Security
Market Equilibrium, Journal of Financial and Quantitative Analysis 20, 407-422.

Berger, Allen N. , Marco A. Espinosa-Vega, W. Scott Frame, and Nathan H. Miller,

2005, Debt Maturity, Risk, and Asymmetric Information, Journal of Finance 60,
2895-2923.

Bharath, Sreedhar T., Paolo Pasquariello, and Guojun Wu, 2009, Does Asymmetric
Information Drive Capital Structure Decisions?, Review of Financial Studies 22,
3211-3043.

Bolton, Patrick,and Mathias Dewatripont, 2005, Contract Theory, The MIT Press, Cam-
bridge, MA.

Booth, L., V. Aivazian, A. Demirguc-Kunt, and V. Maksimovic, 2001, Capital Structures

in Developing Countries, Journal of Finance 56, 87-130.

Botosan, Christine A., and Marlene A. Plumlee, 2005, Assessing Alternative Proxies for
the Expected Risk Premium, The Accounting Review 80, 21-53.
 32

Chordia, Tarun , Richard Roll, and Avanidhar Subrahmanyam, 2000, Commonality in

Liquidity, Journal of Financial Economics 56, 3-28.

Coles, Jeffrey L., Uri Loewenstein, and Jose Suay, On Equilibrium Pricing under Param-
eter Uncertainty, 1995, Journal of Financial and Quantitative Analysis 30, 347-364.

Demirgüc-Kunt, Asli, and Ross Levine, 2001, Bank-Based and Market-Based Finan-
cial Systems: Cross-Country Comparisons, in Financial Structure and Economic
Growth: A Cross-Country Comparison of Banks, Markets, and Development, MIT
Press, Cambridge, MA.

Demirgüc-Kunt, Asli, and Vojislav Maksimovic, 1998, Law, Finance, and Firm growth,
Journal of Finance 53, 2107-2137.

Djankov, S., O. Hart, C. McLiesh; and A. Shleifer, 2008, Debt Enforcement around the
World, Journal of Political Economy 116, 11051149.

Easley, David, Nicholas M. Kiefer, and Maureen O’Hara, 1996, Cream-skimming or

Profit-sharing? the Curious Role of Purchased Order Flow, Journal of Finance 51,
811-833.

Easley, David, and Maureen O’Hara, 2004, Information and the Cost of Capital, Journal
of Finance 59, 1553-1583.

Easley, David, Soeren Hvidkjaer, and Maureen O’Hara, 2010, Factoring Information into
Returns, Journal of Financial and Quantitative Analysis 45, 293-309.

Easton, Peter, 2004, PE Ratios, PEG Ratios, and Estimating the Implied Expected Rate
of Return on Equity Capital, The Accounting Review 79, 73-95.

Fama, Eugene, and Kenneth R. French, 2002, Testing Trade-Off and Pecking Order
Predictions About Dividends and Debt, Review of Financial Studies 15, 1-33.
 33

Fama, Eugene F., 1985, What’s Different About Banks?, Journal of Monetary Economics
15, 29-39.

Fama Eugene F., and James D. MacBeth, 1973, Risk, Return, and Equilibrium: Empir-
ical Tests, Journal of Political Economy 81, 67-105.

Fan, Joseph P.H., Sheridan Titman, and Garry Twite, 2010, An International Com-
parison of Capital Structure and Debt Maturity Choices, Journal of Financial and
Quantitative Analysis, Forthcoming.

Flannery, Mark, 1986, Asymmetric Information and Risky Debt Maturity Choice, The
Journal of Finance 41, 19-37.

Francis, Jere R., Inder K. Khurana, and Raynolde Pereira, 2005, Disclosure Incentives
and Effects on Cost of Capital around the World, The Accounting Review 80, 1125-
1162.

Frank, Murray Z., and Vidhan K. Goyal, 2003, Testing the Pecking Order Theory of
Capital Structure, Journal of Financial Economics 67, 217-248.

Frieder, Laura, and Rodolfo Martell, 2006, On Capital Structure and the Liquidity of a
Firm’s Stock, Working Paper at Purdue University.

Giannetti, M., 2003, Do Better Institutions Mitigate Agency Problems? Evidence from
Corporate Finance Choices, Journal of Financial and Quantitative Analysis 38, 185-
212.

Glosten, Lawrence R., 1987, Components of the Bid-ask Spread and the Statistical
Properties of Transaction Prices, Journal of Finance 42, 1293-1307.

Goel, Anand, and Anjan V. Thakor, 2003, Why Do Firms Smooth Earnings?, Journal
of Business 76, 151-192.
 34

Grossman, Sanford and Joseph Stiglitz, 1980, On the Impossibility of Informationally

Efficient Markets, American Economic Review 70, 393-408.

Hail, Luzi, and Christian Leuz, 2006, International Differences in the Cost of Equity
Capital: Do Legal Institutions and Securities Regulation Matter?, Journal of Ac-
counting Research 44, 485-531.

Horton, Joanne, George Serafeim, and Ioanna Serafeim, 2008, Does Mandatory IFRS
Adoption Improve the Information Environment, Contemporary Accounting Re-
search, Forthcoming.

Johnson, Shane A., 2003, Debt maturity and the effects of growth opportunities and
liquidity risk on leverage, Review of Financial Studies 16, 209-236.

Jung, Kooyu, Yong-Cheol Kim, and Rene M. Stulz, 1996, Timing, Investment Opportu-
nities, Managerial Discretion, and the Security Issue Decision, Journal of Financial
Economics 42, 159-185.

La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer, and Robert Vishny, 1998,
Law and Finance, Journal of Political Economy 106, 1113-1155.

La Porta, Rafael, Florencio Lopez-de-Silanes, and Andrei Shleifer, 2006, What Works in
Securities Laws?, Journal of Finance 61, 1-32.

Lambert, Richard, Christian Leuz, and Robert E. Verrecchia, 2007, Accounting Infor-
mation, Disclosure, and the Cost of Capital, Journal of Accounting Research 45,
385-420.

Lang, Mark H., and Russell J. Lundholm, 1996, Corporate Disclosure Policy and Analyst
Behavior, The Accounting Review 71, 467-492.

Leary, Mark T., and Michael R.Roberts, 2010, The Pecking Order, Debt Capacity, and
Information Asymmetry, Journal of Financial Economics 95, 332-355.
 35

Leuz, Christian , Dhananjay Nandab, and Peter D. Wysocki, 2003, Earnings Manage-
ment and Investor Protection: An International Comparison, Journal of Financial
Economics 69, 505-527.

Mansi, Sattar A., William F. Maxwell, and Darius P. Miller, 2011, Analyst Forecast
Characteristics and the Cost of Debt, Review of Accounting Studies 16, 116-142.

Myers, Stewart C., and Nicholal S. Majluf, 1984, Corporate Financing and Investment
Decisions When Firms Have Information that Investors Do Not Have, Journal of
Financial Economics 13, 187-221.

Myers, Stewart C., 1977, Determinants of corporate borrowing, Journal of Financial

Economics 5,147-175.

Shyam-Sunder, Lakshmi, and Stewart C. Myers. 1999, Testing Static Tradeoff Against
Pecking Order Models of Capital Structure, Journal of Financial Economics 51,
219-244.

Stohs, Mark Hoven, and David C. Mauer, 1996, The determinants of corporate debt
maturity structure, Journal of Business 69, 279-312.

Rajan, Raghuram, and Luigi Zingales, 1995, What Do We Know about Capital Struc-
ture? Some Evidence from International Data, The Journal of Finance 50, 1421-
1460.
 36

TABLE 1
Capital Structure and Cost of Capital Sorted by Information Asymmetry Index
This table presents the univariate tests of the relation between the extent of information asymmetry and capital
structure and the cost of capital. DebtM is market leverage. LT D is long-term debt ratio. COD is the cost of
debt capital. COE is the implied cost of equity capital. Decile 1 represents the group of firm-year observations
with the lowest information asymmetry. Decile 10 represents the group of firm-year observations with the highest
information asymmetry. 10-1 is the difference in mean values between decile 10 and decile 1. t stat is the
t−statistics for the difference between decile 10 and decile 1. The sample period is from 1997 to 2007.

1 2 3 4 5 6 7 8 9 10 10-1 t stat
DebtM -0.0248 -0.0171 -0.0090 -0.0105 -0.0043 0.0052 0.0045 0.0082 0.0119 0.0355 0.0603 25.45
LT D 0.0826 0.0449 0.0252 0.0126 0.0086 -0.0053 -0.0086 -0.0360 -0.0546 -0.0769 -0.1595 -36.85
COD -0.0028 -0.0005 0.0041 0.0001 0.0044 -0.0002 0.0026 -0.0034 -0.0075 0.0031 0.0058 1.11
COE -0.0196 -0.0148 -0.0084 -0.0035 0.0001 0.0050 0.0067 0.0247 0.0412 0.0611 0.0807 20.52
Nobs 8947 8948 8948 8948 8947 8137 9759 8948 8948 8947
 37

TABLE 2
Effects of Information Asymmetry on Market Leverage
This table shows the regressions of firms’ market leverage on the information asymmetry index . Inf Asy is
the information asymmetry index. DebtM is market leverage. F A is tangibility ratio. Size is the logarithm
of total assets. ROA is the return on total assets. M B is the market to book ratio. Inf l is the inflation rate
in percentage. rlend is the lending rate. Bank is the domestic credit provided by banking sector over GDP.
Law is the legal origin dummy. CIF AR measures the level of accounting disclosure practice. All variables
are detailed in Appendix A and firm-level variables are winsorized at 1% and 99% level. t−statistics are
computed based on clustered standard errors at firm level. Year and industry fixed effects are included.
N obs is number of observations. R̄2 is the adjusted R2 . Coefficients with at least 10% significance level are
displayed in bold. The sample period is from 1997 to 2007.
Dependent Variable = DebtM
Full Sample Full Sample Ex. Japan 97-01 02-07 90-07 FM Diff Tobit
1 2 3 4 5 6 7 8 9
Inf Asy 0.023 0.016 0.014 0.014 0.015 0.018 (0.017) 0.008 0.057
(21.71) (14.95) (11.91) (9.73) (14.71) (12.27) (9.30) (17.23) (36.21)
FA 0.158 0.175 0.152 0.172 0.172 0.178 0.172 0.14 0.697
(22.30) (22.71) (19.48) (14.19) (23.35) (22.76) (28.20) (20.14) (52.02)
Size 0.027 0.013 0.015 0.014 0.013 0.010 0.013 0.057 0.062
(28.81) (19.21) (19.27) (15.23) (19.40) (16.16) (22.83) (16.99) (44.37)
ROA -0.128 -0.094 -0.086 -0.233 -0.065 -0.106 -0.166 -0.071 -0.292
-(22.22) (-16.74) (-15.79) (-15.66) (-12.61) (-18.53) (-3.67) -(18.00) -(27.95)
MB -0.012 -0.012 -0.012 -0.027 -0.009 -0.013 -0.018 -0.009 -0.021
-(4.74) (-4.65) (-4.51) (-14.58) (-4.09) (-4.70) (-5.61) -(2.49) -(32.67)
Bank 0.009 0.031 0.021 0.004 0.009 0.015 0.034
(7.20) (10.95) (10.84) (3.35) (6.90) (3.01) (17.59)
Law 0.024 0.015 0.052 0.013 0.017 0.034 0.055
(6.78) (3.87) (9.26) (3.77) (4.68) (3.31) (10.36)
CIF AR -0.335 -0.366 -0.491 -0.259 -0.332 -0.377 -1.419
(-15.13) (-16.37) (-14.41) (-12.47) (-14.90) (-7.38) -(39.43)
Constant -0.093 0.234 0.216 0.409 0.172 0.275 0.286 -0.015 -0.235
-(2.60) (5.22) (4.99) (7.56) (3.76) (6.50) (7.04) -(3.09) -(7.73)
Country FE Yes No No No No Yes No Yes No
Industry FE Yes Yes Yes Yes Yes Yes Yes Yes Yes
Year FE Yes Yes Yes Yes Yes Yes Yes Yes Yes
Nobs 90,514 82,637 63,737 24,751 57,886 82,637 11,710 76,553 82,637
R̄2 28.1% 24.24% 23.93% 27.42% 22.83% 23.72% 25.24% 12.5%
 38

TABLE 3
Effects of Information Asymmetry on Debt Maturity
This table shows the regressions of firms’ debt maturity on the information asymmetry index. Inf Asy
is the information asymmetry index. LT D is long-term debt ratio. F A is tangibility ratio. Size is the
logarithm of total assets. ROA is the return on total assets. M B is the market to book ratio. Bank is the
domestic credit provided by banking sector over GDP. Law is the legal origin dummy. CIF AR measures
the level of accounting disclosure practice. All variables are detailed in Appendix A and firm-level variables
are winsorized at 1% and 99% level. t−statistics are computed based on clustered standard errors at firm
level. N obs is number of observations. R̄2 is the adjusted R2 . Coefficients with at least 10% significance
level are displayed in bold. The sample period is from 1997 to 2007.
Dependent Variable = LT D
Full Sample Full Sample Ex. Japan 97-01 02-07 90-07 FM Diff Tobit
1 2 3 4 5 6 7 8 9
Inf Asy -0.009 -0.039 -0.034 -0.042 -0.038 -0.023 -0.039 -0.002 -0.850
-(6.00) (-25.00) (-17.59) (-18.87) (-21.04) (-10.82) (-23.12) -(1.63) -(49.30)
F A/T A 0.271 0.237 0.228 0.299 0.212 0.224 0.247 0.194 6.932
(25.49) (18.41) (16.82) (15.51) (15.59) (17.00) (17.36) (9.60) (40.18)
Size 0.039 -0.015 -0.001 -0.025 -0.009 -0.006 -0.016 0.061 -0.329
(27.00) (-11.63) (-0.99) (-15.50) (-6.68) (-4.48) (-5.11) (11.45) -(26.61)
ROA -0.006 0.083 0.056 0.153 0.075 0.128 0.105 0.043 2.259
-(0.53) (6.99) (4.78) (5.88) (6.14) (10.44) (6.35) (4.70) (16.73)
MB 0.425 0.005 0.004 -0.005 0.012 0.011 0.004 0.276 0.169
(2.45) (2.68) (2.12) (-2.32) (5.17) (5.35) (1.34) (1.74) (8.57)
Bank 0.002 0.134 -0.006 0.004 0.003 0.004 0.016
(0.82) (23.38) (-1.86) (1.76) (1.30) (0.65) (0.67)
Law 0.054 -0.002 0.025 0.063 0.075 0.048 1.336
(8.31) (-0.27) (2.74) (9.49) (11.71) (7.61) (18.00)
CIF AR 0.430 0.244 0.626 0.356 0.424 0.455 11.245
(10.81) (5.74) (10.85) (8.75) (10.69) (9.46) (24.95)
Constant 0.422 0.407 0.191 0.378 0.398 0.298 0.407 -0.023 -7.348
(4.50) (5.61) (2.39) (3.81) (6.74) (3.96) (24.87) -(2.52) -(18.24)
Nobs 82,069 74,589 56,958 23,353 51,236 74,589 10,485 68,254 74,589
R̄2 31.8% 13.41% 15.23% 17.52% 12.32% 11.44% 14.68% 0.9%
Country FE Yes No No No No No No Yes No
Industry FE Yes Yes Yes Yes Yes Yes Yes Yes Yes
Year FE Yes Yes Yes Yes Yes Yes Yes Yes Yes
 39

TABLE 4
Information Asymmetry and Capital Structure Choices
This table shows the regressions of firms’ capital choices on the information asymmetry index. Inf Asy is
the information asymmetry index. StockIss is stock issuance and DebtIss is debt issuance. Level represents
total dollar amount of securities issued. Dummy equals 1 if debt issuance is above zero or equity issuance
is above 5%, and equals 0 otherwise. F A is tangibility ratio. Size is the logarithm of total assets. ROA
is the return on total assets. M B is the market to book ratio. Bank is the domestic credit provided by
banking sector over GDP. Law is the legal origin dummy. CIF AR measures the level of accounting disclosure
practice. All variables are detailed in Appendix A and firm-level variables are winsorized at 1% and 99%
level. t−statistics are computed based on clustered standard errors at firm level. Year and industry fixed
effects are included. N obs is number of observations. R̄2 is the adjusted R2 . Coefficients with at least 10%
significance level are displayed in bold. The sample period is from 1997 to 2007.
StockIss DebtIss
Level Dummy Level Dummy

Inf Asy -0.005 -0.005 -0.173 -0.188 0.009 0.010 0.172 0.190
-(3.46) -(3.63) -(14.66) -(15.58) (1.82) (1.98) (12.16) (13.14)
FA 0.005 0.006 -0.708 -0.707 0.005 0.010 1.137 1.131
(1.09) (1.20) -(12.54) -(12.29) (0.16) (0.30) (18.10) (17.83)
Size -0.017 -0.017 -0.027 -0.033 0.004 0.005 0.391 0.397
-(15.06) -(15.16) -(3.26) -(3.95) (1.25) (1.40) (37.32) (37.44)
ROA -0.171 -0.171 -0.433 -0.462 -0.017 -0.016 -0.125 -0.112
-(17.40) -(17.20) -(5.14) -(5.40) -(0.55) -(0.51) -(2.23) -(1.98)
MB 0.009 0.009 0.272 0.266 0.005 0.005 -0.110 -0.106
(1.88) (1.83) (20.33) (19.59) (0.91) (0.88) -(11.81) -(11.28)
Bank 0.007 -0.028 0.001 0.154
(7.98) -(1.04) (0.26) (4.60)
Law -0.020 0.009 -0.010 -0.324
-(2.56) (0.07) -(0.61) -(1.94)
CIF AR 0.137 5.681 -0.121 -1.662
(4.75) (6.91) -(2.90) -(1.39)
Constant 0.265 0.155 0.125 -4.010 -0.002 0.108 -3.167 -2.167
(10.42) (10.11) (1.10) -(6.09) -(0.05) (1.85) -(22.67) -(2.27)

Country FE Yes No Yes No Yes No Yes No

Industry FE Yes Yes Yes Yes Yes Yes Yes Yes
Year FE Yes Yes Yes Yes Yes Yes Yes Yes
Nobs 41,600 39,659 41,600 39,659 41,770 39,835 41,770 39,835
R̄2 34.5% 35.0% 22.1% 20.2% 0.7% 0.5% 19.5% 19.5%
 40

TABLE 5
Effects of Information Asymmetry on the Cost of Capital
This table shows the regressions of firms’ market leverage, cost of equity and cost of debt capital on the
information asymmetry index. Inf Asy is the information asymmetry index. COD is the cost of debt capital.
COE is the implied cost of equity capital. CODstd − COEstd is the difference between standardized COD
and standardized COE. F A is tangibility ratio. Size is the logarithm of total assets. ROA is the return
on total assets. M B is the market to book ratio. DebtB is book leverage. Beta is market beta from June
of year t. F Bias is the forecasting bias. Inf l is the inflation rate in percentage. rlend is the lending rate.
Bank is the domestic credit provided by banking sector over GDP. Law is the legal origin dummy. CIF AR
measures the level of accounting disclosure practice. All variables are detailed in Appendix A and firm-level
variables are winsorized at the 1% and 99% level. t−statistics are computed based on clustered standard
errors at firm level. Year and industry fixed effects are included. N obs is number of observations. R̄2 is the
adjusted R2 . Coefficients with at least 10% significance level are displayed in bold. The sample period is
from 1997 to 2007.

COD COE CODstd − COEstd

OLS Std Reg OLS OLS Std Reg OLS OLS OLS
1 2 3 4 5 6 7 8

Inf Asy 0.004 0.002 0.006 0.010 0.085 0.018 -0.115 -0.155
(0.30) (0.53) (10.42) (17.69) -(7.52) -(10.31)
F A/T A 0.130 0.008 0.185 -0.096 -0.119
(1.04) (1.38) -(1.47) -(1.81)
Size -0.005 -0.003 0.009 -0.014 -0.377 -0.001 0.080 0.024
-(0.50) (0.50) -(26.67) -(2.22) (9.15) (4.34)
ROA -0.113 -0.006 -0.041 1.157 1.206
-(2.59) -(1.50) (9.73) (10.03)
MB 0.027 0.010 0.033 -0.008 -0.102 -0.007 0.096 0.096
(1.78) (1.25) -(11.78) -(9.47) (8.33) (8.22)
DebtB -0.430 -0.031 -0.285 0.128 0.221 0.083 -2.166 -2.088
-(2.37) -(1.75) (26.02) (16.84) -(24.99) -(23.18)
Beta 0.004 0.359 0.001 -1.150 -0.001
(2.39) (3.13) -(0.86) -(0.40)
F bias 0.005 11.841 0.002 0.006 -0.094
(1.77) (1.23) (0.13) -(1.50)
Inf l 0.001 0.104 0.002 0.010 0.036
(6.76) (10.31) (2.25) (5.03)
rLend -1.343 -0.018 12.510 0.188 -1.226
-(0.25) (1.31) (0.31) -(2.89)
Bank 0.098 -0.014 0.033
(0.92) -(22.18) (2.51)
Law -0.387 -0.003 0.027
-(1.48) -(1.62) (0.85)
CIF AR 0.645 -0.028 -0.274
(1.12) -(2.31) -(1.40)
Constant 0.258 -1.193 0.180 0.143 -0.196 0.417
(0.80) -(0.99) (28.79) (8.03) -(1.02) (1.57)
Country FE Yes Yes No Yes Yes No Yes No
Industry FE Yes Yes Yes Yes Yes Yes Yes Yes
Year FE Yes Yes Yes Yes Yes Yes Yes Yes
Nobs 75,595 75,594 57,902 40,076 40,075 38,385 29,839 28,440
barR2 4.8% 4.6% 1.8% 24.1% 25.4% 17.0% 13.3% 12.8%
 41

TABLE 6
Country Infrastructure and Capital Structure
This table shows the effects of country-level factors on the relation between information asymmetry and
market leverage. Inf Asy is the information asymmetry index. DebtM is market leverage. F A is tangibility
ratio. Size is the logarithm of total assets. ROA is the return on total assets. M B is the market to book
ratio. Bank is the domestic credit provided by banking sector over GDP. Law is the legal origin dummy.
CIF AR measures the level of accounting disclosure practice. All variables are detailed in Appendix A
and firm-level variables are winsorized at 1% and 99% level. t−statistics are computed based on clustered
standard errors at firm level. Year and industry fixed effects are included. All regressions include unreported
firm and country characteristics. N obs is number of observations. R̄2 is the adjusted R2 . Coefficients with
at least 10% significance level are displayed in bold. The sample period is from 1997 to 2007.

Dependent Variable = DebtM

1 2 3 4

Inf Asy 0.014 0.023 0.040 0.038

(8.08) (17.24) (4.00) (3.69)
Inf Asy × Bank 0.004 0.004
(5.33) (5.79)
Inf Asy × Law -0.004 0.002
(-2.09) (0.99)
Inf Asy × CIF AR -0.023 -0.033
(-1.84) (-2.47)
FA 0.175 0.175 0.175 0.175
(22.07) (22.07) (22.04) (22.02)
Size 0.026 0.026 0.026 0.026
(26.14) (26.09) (25.76) (25.59)
ROA -0.114 -0.113 -0.112 -0.115
(-20.03) (-19.73) (-19.75) (-19.90)
MB -0.011 -0.011 -0.011 -0.011
(-4.54) (-4.60) (-4.60) (-4.55)
Bank 0.001 0.001 0.001 0.001
(0.49) (0.42) (0.45) (0.56)
Law 0.017 0.017 0.017 0.017
(4.64) (4.62) (4.67) (4.61)
CIF AR -0.319 -0.319 -0.320 -0.321
(-14.35) (-14.41) (-14.45) (-14.44)
Constant 0.008 0.010 0.012 0.015
(0.11) (0.16) (0.18) (0.22)
Country FE Yes Yes Yes Yes
Industry FE Yes Yes Yes Yes
Year FE Yes Yes Yes Yes
Nobs 73,172 73,172 73,172 73,172
R̄2 26.2% 26.1% 26.05% 26.22%
 42

TABLE 7
Adoption of IFRS on Capital Structure and Cost of Capital
This table shows how the adoption of IFRS would affect firms’ capital structure and the cost of cap-
ital. IFRS adopted countries include Australia, Austria, Belgium, Denmark, Finland, France, Ger-
many, Greece, Hong Kong, Ireland, Italy, Netherlands, Norway, Portugal, Spain, Sweden, Switzer-
land, South Africa, and U.K.. Panel A presents the results with IFRS dummy which equals 1 for
observations in years 2007 and thereafter, and 0 for observations in years 2004 and before. Panel B
presents the regressions of the differences between 2007 and 2004 around IFRS adoption in DebtM ,
LT D, COD, COE, on the differences in the information asymmetry index and other control vari-
ables. The differences are computed over the event window of [2004, 2007], i.e., one year before
and two years after the IFRS adoption in 2005. DebtM is market leverage. LT D is long-term debt
ratio. COD is the cost of debt capital. COE is the implied cost of equity capital. F A is tangibility
ratio. Size is the logarithm of total assets. ROA is the return on total assets. M B is the market
to book ratio. DebtB is book leverage. Beta is market beta from June of year t. F Bias is the
forecasting bias. rlend is the lending rate. Inf l is the inflation rate in percentage. All variables are
detailed in Appendix A and firm-level variables are winsorized at 1% and 99% level. t−statistics are
computed based on clustered standard errors at firm level. Country, year and industry fixed effects
are included. N obs is number of observations. R̄2 is the adjusted R2 . Coefficients with at least 10%
significance level are displayed in bold. The sample period is from 1997 to 2007.

Panel A: IFRS Dummy Panel B: Differences around IFRS Adoption

DebtM LT D COD COE DebtM LT D COD COE
1 2 3 4 5 6 7 8
Inf Asy 0.017 -0.011 0.039 0.015 0.020 -0.015 0.004 0.020
(10.53) (-3.84) (1.01) (7.80) (8.64) (-2.03) (1.70) (2.07)
IF RS -0.004 0.014 0.003 -0.006
(-1.95) (2.13) (0.46) (-1.26)
IF RS × Inf Asy -0.003 0.002 -0.037 -0.010
(-1.70) (0.36) (-1.14) (-2.60)
FA 0.132 0.249 0.114 0.101 0.067 0.002
(13.22) (13.19) (0.67) (5.94) (1.12) (0.11)
Size 0.022 0.037 -0.002 -0.016 0.053 0.021 -0.001 -0.001
(19.43) (15.94) (-0.28) (-17.26) (13.21) (1.50) (-0.25) (-0.14)
ROA -0.058 -0.011 -0.015 -0.042 0.048 -0.018
(-11.83) (-0.78) (-0.84) (-5.93) (1.57) (-1.17)
MB -0.014 0.006 0.011 -0.009 -0.002 0.005 0.003 -0.010
(-14.15) (2.51) (1.50) (-9.18) (-1.61) (0.80) (1.15) (-1.88)
DebtB -0.173 0.128 -0.116 0.129
(-1.40) (13.24) (-5.45) (3.37)
Beta 0.034 -0.320
(1.99) (-0.94)
F bias 0.000 0.001
(2.53) (2.73)
Inf l 0.001 1.603
(1.99) (1.77)
rLend 0.487 0.013
(1.66) (0.87)
Constant 0.066 0.464 -0.179 0.232 -0.237 0.414 -0.020 -0.004
(0.33) (2.41) (-1.07) (18.45) (-18.62) (2.11) (-2.10) (-0.14)
Nobs 22,709 20,184 14,494 10,676 3,473 2,791 1,983 748
R̄2 28.5% 20.4% 11.0% 25.8% 17.8% 4.8% 8.2% 18.7%
 TABLE 8
Effects of Analyst Activities and Earnings Management on Capital Structure
This table shows regressions of capital structure and debt maturity on alternative information proxies created from analyst activities(Analyst
Variales) and earnings management(Earnings MGMT). # of Analyst is the number of analysts covering a firm. F disp is analyst forecast
dispersion. F Err is analyst forecast errors. Accrual is the magnitude of accruals divided by operating cash flows. ESmooth is earnings
smoothing over the last 5 years. ECorr is the earnings correlation over the last 5 years. DebtM is market leverage. LT D is long-term debt
ratio. F A is tangibility ratio. Size is the logarithm of total assets. ROA is the return on assets. M B is the market to book ratio. Bank
is the domestic credit provided by banking sector over GDP. Law is the legal origin dummy. CIF AR measures the level of accounting
disclosure practice. All variables are detailed in Appendix A and firm-level variables are winsorized at 1% and 99% level. t−statistics are
computed based on clustered standard errors at firm level. Year and industry fixed effects are included. N obs is number of observations.
R̄2 is the adjusted R2 . Coefficients with at least 10% significance level are displayed in bold. The sample period is from 1997 to 2007.

Dependent Variable = DebtM Dependent Variable = LT D

Analyst Variables Earnings MGMT Analyst Variables Earnings MGMT
# of Analyst F disp F Err Accrual ESmooth ECorr # of Analyst F disp F Err Accrul ESmooth ECorr
1 2 3 4 5 6 7 8 9 10 11 12

Analyst -0.002 0.003 0.000 0.013 -0.001 0.000

(-9.28) (1.78) (1.79) (27.45) (-1.39) (-4.16)
EM GM T 0.021 -0.014 -0.040 -0.044 0.004 0.014
(2.35) (-4.33) (-12.16) (-1.96) (1.52) (2.03)
FA 0.155 0.134 0.154 0.178 0.180 0.183 0.230 0.222 0.232 0.221 0.217 0.210
(17.01) (13.26) (16.71) (21.71) (20.72) (20.34) (14.78) (12.28) (14.00) (16.26) (15.28) (14.24)
Size 0.008 0.006 0.006 0.009 0.008 0.007 -0.023 -0.016 -0.014 -0.004 -0.004 -0.005
(11.29) (8.22) (9.12) (13.62) (11.87) (10.59) (-18.01) (-11.84) (-10.37) (-3.17) (-3.53) (-3.73)
ROA -0.160 -0.194 -0.170 -0.109 -0.125 -0.145 -0.016 -0.009 0.058 0.127 0.151 0.154
(-17.19) (-14.56) (-17.74) (-18.61) (-19.59) (-20.21) (-0.85) (-0.38) (3.13) (10.01) (11.51) (10.85)
MB -0.029 -0.031 -0.031 -0.016 -0.016 -0.018 -0.015 -0.009 -0.002 0.013 0.016 0.018
(-20.65) (-21.03) (-21.16) (-4.61) (-4.13) (-4.57) (-7.05) (-3.42) (-0.87) (6.32) (7.22) (7.38)
Bank 0.007 0.004 0.007 0.008 0.006 0.006 -0.005 -0.001 -0.004 0.002 -0.001 -0.002
(5.06) (2.33) (4.83) (6.44) (4.82) (4.28) (-2.21) (-0.46) (-1.57) (0.74) (-0.32) (-1.09)
Law 0.017 0.015 0.009 0.014 0.013 0.011 0.043 0.085 0.078 0.077 0.080 0.084
(4.24) (3.66) (2.39) (3.81) (3.46) (2.91) (5.50) (9.97) (9.97) (11.83) (12.04) (12.23)
CIF AR -0.214 -0.180 -0.209 -0.322 -0.323 -0.325 0.466 0.527 0.489 0.429 0.444 0.461
(-9.48) (-7.42) (-9.16) (-14.56) (-14.02) (-13.53) (10.00) (10.04) (10.09) (10.74) (10.71) (10.67)
Constant 0.278 0.300 0.286 0.312 0.343 0.314 0.419 0.308 0.349 0.220 0.186 0.177
(4.52) (5.53) (5.33) (6.47) (6.86) (6.26) (2.62) (3.02) (3.19) (2.87) (2.39) (2.22)
Country FE No No No No No No No No No No No No
Industry FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Year FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Nobs 50,337 39,550 49,547 80,434 73,847 67,851 46,846 37,098 46,142 72,693 67,122 61,962
43

R̄2 27.40% 26.90% 26.95% 23.49% 23.79% 24.14% 18.68% 15.35% 14.01% 11.16% 11.81% 12.50%