Exodus From Sovereign Risk

Exodus from Sovereign Risk:
Global Asset and Information Networks in the Pricing of Corporate
Credit Risk
Jongsub Lee, Andy Naranjo, and Stace Sirmans
Department of Finance, Insurance & Real Estate

Warrington College of Business Administration, University of Florida
August 29, 2014
Abstract
Using 5-year credit default swap (CDS) spreads on 2,364 companies in 54 countries during
2004-2011, we show that firms exposed to better property rights institutions through their for-
eign asset positions (Institutional channel) and firms whose stocks are listed on exchanges with
stricter disclosure requirements (Informational channel) reduce their CDS spreads by 40 bps
for a one standard deviation increase in their exposure on the two channels. These channels
capture distinct effects beyond those associated with firm- and country-level fundamentals.
Overall, we find that firm-level global asset and information connections are important mech-
anisms to delink firms from their sovereign risk.
JEL Classification: F23, F34, F36, G01, G15, K40

Keywords: International CDS markets, corporate credit risk, sovereign risk, multinational
corporations, asset geographic location, cross-listing, property rights, creditor rights, disclosure
requirements, sovereign ceiling violations

We thank Viral Acharya, David T. Brown, Mark Flannery, Ken Singleton (the Editor), Francis Longstaff, Jay
Ritter, Marti Subrahmanyam, and an anonymous Associate Editor as well as two referees for helpful comments and
suggestions. We also thank both the Heritage Foundation, particularly Anthony B. Kim, and the Credit Research
group at Markit led by Gavan Nolan for their data support. We are grateful for the many insightful comments from
our discussants and participants at the 2013 SFS Finance Cavalcade, the 2013 WU Gutmann Symposium, and the
2013 China International Conference in China. Jaeyung Kim and Jerry Singh provided excellent research assistance.

Assistant Professor, Tel: 352.273.4966, Email: jongsub.lee@warrington.ufl.edu

Bank of America Associate Professor and Director of CIBER, Tel: 352.392.3781, Email:
andy.naranjo@warrington.ufl.edu

Ph.D. Candidate, Tel: 352.459.2039, Email: stace.sirmans@warrington.ufl.edu, Website: www.stacesirmans.com
Electronic copy available at: http://ssrn.com/abstract=2492204

Exodus from Sovereign Risk Page 1
To what extent can private sector firms delink themselves from sovereign risk? It is well known that
while governments play an important role in financial markets, they can also generate externalities
arising from their actions and their ability to compel and proscribe. For example, governments
suffering from large budget deficits and slow economic growth may limit the ability of private
sector firms to service their external debt obligations. Budget deficits may also presage increases
in corporate taxes or other corporate infringements, increasing potential default risks. 1 In the
aftermath of the recent global credit risk crisis where sovereign credit default swap (CDS) spreads
rose dramatically from an average of 30 bps to above 250 bps in 2008 and remained above 100 bps
thereafter (see Figure 1), there is an increasing concern about the transferring of sovereign risks to
private sector firms. Anecdotal evidence further suggests that these corporate and investor concerns
are warranted, as governments such as Argentina have turned towards heterodox, interventionist
economic policies whereby they have imposed increasingly invasive corporate restrictive measures
such as expropriations and foreign-exchange, trade, and capital controls. 2 Major credit rating
agencies term such investor concerns transfer and convertibility (T&C) risk - the risk of exchange
controls being imposed by sovereign authorities that prevent or impede the private sectors ability
to convert local currency into foreign currency and make their debt service. 3
[ Insert Figure 1 here ]
In this paper, we propose two novel institutional and informational channels through which
private sector firms can, in part, delink themselves from their sovereign risk. 4 Using 5-year CDS
spreads on 2,364 companies in 54 countries during 2004-2011, we show that firms exposed to better
property rights institutions through their foreign asset positions (Institutional channel) and firms
whose stocks are listed on exchanges with stricter disclosure requirements (Informational channel)
reduce their CDS spreads by 40 bps for a one standard deviation increase in their exposures on the
two channels. Our reported effects exist above and beyond those associated with both firm- and
sovereign-level fundamentals as well as credit rating effects. We further show that our results are
1
See, for example, Ernst and Young (January, 2012), Global Tax Policy and Controversy Briefing.
2
See, for example, Wall Street Journal : Argentina to Seize Control of Oil Firm (M. Moffett and T. Turner,
April 17, 2012) and Dollars Become Scarce as Argentina Cries Peso (M. Moffett, June 13, 2012).
3
More broadly, these T&C type risks also include government expropriatory actions such as nationalizing firm
assets or taking actions that restrict the ability of a company to operate. See, for instance, S&P (2008), 2008
Corporate Criteria: Analytical Methodology.
4
We interchangeably use the terms, T&C risk, sovereign risk, and sovereign governments grabbing-hand risk,
throughout the text hereafter.
Electronic copy available at: http://ssrn.com/abstract=2492204

robust to extensive robustness checks, including potential specification, identification, and endo-
genity concerns. Overall, our results highlight that firm-level global asset and information network
connections are important mechanisms to delink firms from their sovereign risk.
Our study contributes to three important research streams. First, we contribute to the literature
on multinational corporate credit risk. Our international CDS data and the information on both
firm- and sovereign-level fundamentals provide a comprehensive picture on the determinants of
corporate credit risk for 2,364 firms in 54 countries over an 8-year period. 5 We extend the credit
risk models explored in Collin-Dufresne, Goldstein, and Martin (2001), Bharath and Shumway
(2008), and Campbell, Hilscher, and Szilagyi (2008) by formally testing a more elaborate credit risk
pricing model for multinational corporations (MNCs) whose global asset and information network
connections play an important role in explaining their borrowing costs. International CDS markets
are an ideal experimental setting to conduct these tests relative to corporate bond yield spreads
and credit ratings because CDS contracts are standardized in their contractual terms and are also
more liquidly traded than the bonds issued by the same entities (Longstaff, Mithal, and Neis, 2005).
Moreover, CDS investors are often perceived as more sophisticated and informed than investors in
other financial markets (Acharya and Johnson, 2007; Berndt and Ostrovnaya, 2008), which suggests
potentially greater market information content than credit ratings ( Hull, Predescu, and White, 2004;
Flannery, Houston, and Partnoy, 2010; Chava, Ganduri, and Ornthanalai, 2012).
Second, we contribute to the international corporate linkages and exposure literature by uniquely
highlighting the role of a firms asset distribution across countries and the firms equity cross-listing
status in determining the firms effective exposure to institutional and informational environments
across countries. Several studies in this literature highlight how the foreign affiliates of MNCs bring
international exposures to the MNCs capital structure through the effects of varying tax rates
and creditor rights in the countries of their foreign affiliates (Noe, 1998; Desai, Foley, and Hines,
2004). The literature also emphasizes that stricter disclosure of information through equity cross-
listings in more transparent markets helps a firm mitigate the information risk borne by the firms
non-local shareholders, thereby enhancing the firms access to external capital markets ( Pagano,
5
Durbin and Ng (2005) study the determinants of yield spreads on only 108 corporate bonds from emerging
markets, and Peter and Grandes (2005) examine similar topics using bond market data for 9 firms in South Africa.
Borensztein, Cowan, and Valenzuela (2013) examine the relation between corporate credit ratings and their sovereign
counterpart using S&P credit ratings on 509 firms in 30 emerging market countries. Bai and Wei (2012) study similar
topics using international CDS data from 30 countries for 3 years from 2008 to 2010.
Roell, and Zechner, 2002; Lins, Strickland, and Zenner, 2005; Bailey, Karolyi, and Salva, 2006).6
Our global asset distribution and equity cross-listing channels extend these international corporate
linkage notions to corporate credit risk pricing.
Third, we contribute to the law and economics literature on the effects of a local governments
legal and institutional characteristics on the structure and pricing of financial claims ( La Porta,
Lopez de Silanes, Shleifer, and Vishny, 1997, 1998, 2002; Shleifer and Wolfenzon, 2002; Qian and
Strahan, 2007; Bae and Goyal, 2009). Our study highlights that not only do the home countrys
institutional characteristics matter for the pricing of a firms claims, but so do those of foreign
institutions that the firm is connected to through global asset and information connections. Im-
portantly, we show that firms can opt out of their legal and institutional heritages by utilizing
these global network connections. In this regard, our study extends the Coasian view on investor
protection proposed by Bergman and Nicolaievsky (2007) who explain how private sector entities
overcome legal and institutional externalities through contracts. 7
As some visual motivation, Panel A of Figure 2 shows a dramatic increase in the frequency and
magnitude of incidences where 5-year corporate CDS spreads fall below their sovereign counterpart
with equal contractual terms, implying better credit prices of private sector entities than their
sovereign government. These sovereign ceiling rule violations in international CDS markets
situations in which corporate-sovereign CDS spread differences are strictly negative are especially
pronounced since the onset of the global sovereign credit risk crisis. The magnitude of these CDS
spread differences is non-trivial, reaching over 100 bps per annum, on average. 8 Looking at their
cross-sectional patterns in Panel B of Figure 2, we further find that these negative corporate-
sovereign CDS spread differences are prevalent in countries with weak institutional characteristics
in terms of property rights/creditor rights protections and low information disclosure requirements.
Motivated by these temporal and cross-sectional 5-year corporate-sovereign spread difference
patterns in international CDS markets, we examine how each firms foreign asset geographic lo-
6
For a discussion on countrywide liberalizations, see Bekaert and Harvey (2000) and Bekaert, Harvey, and Lund-
blad (2005).
7
Miller and Reisel (2012) provide evidence on the ability of security-level contracts to overcome the deficiencies of
country-level creditor protections. Nini, Smith, and Sufi (2012) also find a significant influence of contractual credit
rights on corporate investment policy and corporate governance.
8
The average CDS spread difference between AAA- and BBB-rated companies during the same time period was
110 bps.
cation and its stock cross-listing status in foreign markets serve as mechanisms to delink firms
from their sovereign risk. We hypothesize that firms whose foreign assets are located in countries
with better property rights institutions than their home country mitigate their investors T&C risk
concerns due to stronger legal rules and conditions that govern the firms assets in those foreign
markets (Institutional channel). 9 We also conjecture that stricter disclosure requirements imposed
by foreign exchanges where a firms stock is cross-listed alleviate T&C risk concerns through im-
proved transparency and reductions in firm informational frictions (Informational channel). We
expect that the influence of both channels is greater during the recent sovereign credit risk crisis
given the increases in government distress and deterioration of firm operational information. We
also expect that among the two channels, the institutional channel tends to better capture recovery
rate effects in the event of default on the pricing of corporate credit because stronger property rights
and creditor rights protections enhance the efficiency of post-default restructuring processes.
We test our institutional and informational channels in MNC credit risk pricing using the credit
risk models examined by Collin-Dufresne, Goldstein, and Martin (2001), Bharath and Shumway
(2008), and Campbell, Hilscher, and Szilagyi (2008). As proxies for each countrys institutional
quality, we use various property rights protection measures compiled by the Heritage Foundation and
International Country Risk Guide (ICRG) databases as well as creditor rights proxies constructed
by La Porta, Lopez de Silanes, Shleifer, and Vishny (1998) and Djankov, McLiesh, and Shleifer
(2007). We use the required number of disclosed items and disclosure frequency mandated by local
exchanges as proxies for the countrys informational quality (Bushman, Piotroski, and Smith, 2004).
Controlling for a comprehensive list of firm- and sovereign-level credit risk factors, we find that
a one standard deviation increase in a firms foreign asset exposure to stronger property rights and
creditor rights institutions than its home country and exposure to stricter disclosure requirements
mandated by foreign stock exchanges where the firm stocks are cross-listed reduce an average firms
5-year CDS spread by 40 bps. We also find stronger effects from these two channels during the
recent sovereign credit risk crisis. We further show that our institutional channel is more closely
related to recovery rate effects in the pricing of credit. We find a 3.16% enhancement in corporate
CDS recovery rates from their sample average (37.042%) for a one standard deviation increase in a
9
A recent paper by Choi, Gulati, and Posner (2011) examines the spread between the yield of Greek sovereign
bonds that have Greek choice of law terms and that of Greek sovereign bonds that have English choice of law terms.
They find that Greek law bonds have higher yields and lower prices, and the spread between the two types of bonds
increases in probability of Greek default.
firms exposure to better property rights and creditor rights institutions. However, we do not find
significant recovery rate effects associated with our informational channel.
Our results are robust to numerous identification, specification, selection, and reverse causality
tests. Importantly, we show that it is the geographic location of a firms asset that matters; neither
the aggregate amount of a firms foreign assets nor sales explains the corporate CDS spreads. This
identification highlights the importance of a firms foreign asset geographic information in measuring
the MNCs effective institutional exposure. Our results are also robust to controlling for the firms
exposure to foreign fundamentals such as GDPs and stock market volatilities of connected foreign
countries through their foreign asset positions. Moreover, our global asset and information network
connections are not fully incorporated in S&P corporate credit ratings, and therefore our results
carry through even after controlling for credit rating effects. We also show that our institutional
and informational factors explain sovereign ceiling violation (SCVs) patterns in international CDS
markets using alternative model specifications, including analyses with Probit and Ordered-Probit
models.
Arguably, selection and reverse causality are also potential concerns. Instead of our regressions
capturing the effects of better laws and institutions, our institutional and informational channels
could proxy for better quality firms whose operations can go beyond their country of domicile
and place their assets in better institutions than their home markets to reduce borrowing costs.
Nevertheless, we provide several novel tests showing that our results are also robust to selection
and reverse causality concerns further validating that the firms asset location is deterministic in
our documented channel effects.
Overall, our results highlight that firms global asset and information network connections in
countries with better institutional and informational environments are important mechanisms to
delink private sector entities, in part, from their sovereign risk. These results are also important in
the pricing of multinational corporate credit where global network connections play an important
role in borrowing costs.
1 Institutional Background on the CDS Market

A CDS is a bilateral insurance contract over a fixed term whose payment is contingent on the
default of an underlying credit entity (also often called reference entity). The trigger event of
the reference entitys default includes bankruptcy, failure to pay, repudiation and moratorium,
obligation acceleration, obligation default, and restructuring. 10 A CDS contract is associated with
bonds or any credit assets issued by a reference entity, and it trades on the over-the-counter market.
The party that sells the default insurance (protection seller) pays the full face value of the
underlying bond issued by a reference entity upon the entitys default, while the other party that
purchases the insurance claim (protection buyer) periodically pays fees over the term of the swap
to the protection seller until the reference entity defaults or the contract maturity, whichever comes
first. Upon default, the protection buyer delivers the defaulted underlying bond to the protection
seller. With this CDS contract, the buyer transfers the default risk of the reference entity to the
seller, which bears the full loss given default (LGD). The fees collected by the seller are called CDS
spreads, which are quoted in basis points per annum of the contracts notional value and are usually
paid quarterly. The spreads capture both the recovery rate of the underlying bond and the hazard
rate that the reference entity could default in an infinitesimal time interval conditioning that the
entity does not default by that time. 11 The reference entity of the swap could be a private sector
firm or a government. In the former case, we call the CDS contract a corporate CDS, and in the
latter case a sovereign CDS.
Each CDS contract comes with its own insurance period maturity, restructuring clause that
defines trigger events and further denotes the types of bond that a protection buyer could deliver
to a protection seller upon the trigger events, and settlement base currency. Maturities span from
a few months to 10-years or more, but 5-year maturities are the most liquidly traded corporate and
sovereign CDS contracts (Longstaff, Mithal, and Neis, 2005). A restructuring clause determines
the types of trigger events and affects the recovery rate of an underlying obligation by limiting the
types of deliverable obligations upon such events. Therefore, it is possible that two CDS contracts
10
Since the original International Swaps and Derivatives Association (ISDA) agreement in 1999, there are six
general categories of trigger events that mandate the payment from the protection seller to the protection buyer.
They are bankruptcy, failure to pay, repudiation/moratorium, obligation acceleration, obligation default, and re-
structuring. For corporate borrowers, there are three principle credit events such as bankruptcy, failure to pay, and
restructuring. Since 1999, several minor modifications have been made by the ISDA, in April 2001 and January
2003. The modifications are mostly about the definition of restructuring. However, for standard North American
corporate transactions from April 8, 2009 onwards, restructuring events are excluded from trigger events following
2009 ISDA Credit Derivatives Determinations Committees and Auction Settlement CDS Protocol known as Big
Bang Protocol.
11
Under the flat hazard rate and constant recovery rate assumptions, the fair spread can be simplified to the
product of the hazard rate and the constant LGD per $1-notional in a continuous time setup. Hence, the fair spread
captures the expected LGD for a $1-notional under the risk-neutral probability measure. See Internet Appendix E
for more details on this discussion.
with equal contractual terms but different restructuring clauses could come up with different spread
values due to the potential cheapest-to-deliver options (Packer and Zhu, 2005; Berndt, Jarrow,
and Kang, 2007).
There are four distinct restructuring types that are mostly different in delivery options underly-
ing the obligations: 1) Complete restructuring (CR), which allows delivery of any bond of maturity
up to 30 years issued by the underlying credit entity; 2) Modified restructuring (MR), which limits
deliverable obligations to those with a maturity of 30 months or less after the termination date
of the CDS contract; 3) Modified-modified restructuring (MM), which limits deliverable assets to
be bonds with a maturity shorter than 60 months for restructured obligations and 30 months for
all other obligations; and 4) No restructuring (XR), which excludes all restructuring events from
trigger events.
In our analysis, we carefully match the spread of a 5-year corporate CDS to that of its sovereign
counterpart with equal maturity, tier, restructuring clause, and settlement currency to eliminate
any non-credit risk profiles that may cause the two spreads to diverge from each other, even if the
two reference entities are perceived to have similar credit qualities. We also control for various fixed
effect dummies defined at each contractual term to address potential contractual term effects on
CDS spread level.12
2 Hypothesis Development
Suppose that a country is in financial distress and unable to pay back its debt obligations. Consider
a hypothetical firm that has all of its assets located within that country, and thus, the assets are
largely governed by the rules and conditions imposed by the distressed government in which the firm
and its assets reside. One can envision two scenarios for this purely domestic firm, default (D-state)
and non-default (ND-state). In a D-state, the firms creditors may not fully capture the residual
12
However, it should be noted that our contractual term matching and fixed effect controls cannot fully distinguish
the potential differential effects of sovereign CDS spreads on corporate CDS spreads across restructuring clauses, if
any. For a sovereign, the likelihood of a restructuring event is the first order concern in default arrival risk, whereas
for a corporation, it is the likelihood of bankruptcy and failure to pay (see Sovereign CDS: Credit Event and Auction
Primer by Morgan Stanley, 2011). Given that sovereign restructuring may speed up private sector defaults, corporate
CDS spreads paired with sovereign contracts that include restructuring as a part of the trigger events could be more
sensitive to sovereign CDS spreads, other things equal. In Internet Appendix B.2, we run our analysis separately for
different restructuring clause subsamples, including no restructuring (XR) versus the others, and confirm that our
results are robust to these potential sovereign versus corporate CDS trigger event differences.
value of the firms assets if the distressed government intervenes in the post-default restructuring
process. The government can affect the restructuring process either directly by changing the rules
and conditions on how the creditors can pull out the residual value of the defaulted firms assets, or
indirectly by prohibiting any capital outflows from the country. 13 Even in an ND-state, the firms
creditors are still concerned about the increase in the firms default probability that is induced by
potential increases in taxes or other forms of government expropriation attempts on the firms cash
flows or assets. Due to this potential violation of their property rights in either state, lenders are
not willing to pay favorable prices for the firms debt obligations, which results in higher spreads
on CDS contracts written on that firms debt obligations.
However, if the firm had a portion of its assets in other countries with better property rights,
the firms creditors would expect more transparent and credible enforcement of the rules and laws
applied to that firms foreign assets. Alternatively, they may anticipate a potential flow of sub-
sidizing capital from the firms foreign operations to its domestic operations. For these reasons,
the creditors concerns on the local governments T&C risk would be mitigated, which could en-
hance the lenders ex ante expectation on the recovery rate in the event of firm default (recovery
rate effect) and could also lead to a lower perceived default probability of the underlying credit
entity (default probability effect). This could result in a lower firm CDS contract spread. From the
foregoing arguments, we hypothesize:
H1: When a firm has foreign assets in countries with better property rights, the firms
CDS spread is lower, ceteris paribus.
Better creditor rights enhance the efficiency and speed of post-default restructuring and lowers
within-creditor conflict of interest. Creditor rights are particularly relevant when firms are in a
D-state, with better creditor rights increasing the recovery value that creditors can potentially pull
out of debtors in the restructuring process (recovery rate effect). This leads to better ex ante credit
pricing.
H2: When a firm has foreign assets in countries with better creditor rights, the firms
CDS spread is lower, ceteris paribus.
13
This local government indirect capital control could be a particular concern for non-resident claimants.
Without a comprehensive picture of a firms operations, firm investors cannot accurately measure
the firms true default risk (measurement error in default risk forecast) and the potential recovery
value upon the firms default (measurement error in recovery rate forecast). Hence, the price of the
bond issued by a firm whose balance sheet information is limitedly disclosed to its creditors could
be discounted due to these information opacity reasons. Following the notions used in the home
bias literature (Coval and Moskowitz, 2001; Mian, 2006), informational opacity could be a more
serious concern to foreign investors, and such investors would not be willing to pay favorable credit
prices unless the issuer disclosed its accounting information in a more frequent and comprehensive
manner. More stringent equity cross-listing requirements help alleviate foreign investor concerns
since they often require additional degrees of disclosure on the frequency and number of items
regarding corporate balance sheet information. The aforementioned discussion gives rise to our
third hypothesis:
H3: When a firms equity is listed on foreign exchanges with stricter disclosure require-
ments, the firms CDS spread is lower, ceteris paribus.
From the above hypotheses, we expect that firms whose assets are located in countries with
better property rights and creditor rights than those of the firms local government and firms whose
stocks are cross-listed on exchanges with stricter disclosure requirements than its domestic exchange
can better escape their local governments risk transfer attempts. Through these channels, firms
can also possibly decouple themselves to to the point where their credit spreads are lower than their
sovereign counterpart, which are SCVs in international CDS markets.
We expect that our proposed delinking mechanisms work more effectively as the probability
of local governments default increases and, therefore, investors T&C risk concerns become more
severe. We further expect that our institutional channel is more likely to affect recovery rates in the
event of corporate default (D-state) than our informational channel, due in part to the relevance of
the institutional channel in efficient and timely post-default restructuring processes.
3 Data and Variables

We gather daily CDS contract data for corporations and governments over the years 2004-2011 from
Markit, and we use the 5-year CDS contract since it is the most liquidly trading contract among the
CDS maturities (Longstaff, Mithal, and Neis, 2005; Acharya and Johnson, 2007; Ericsson, Jacobs,
and Oviedo, 2009). Markit provides details for each firms CDS contract, including firm-level
information (ticker, country, region, industry, and credit rating composite) and contract-specific
information (tier, restructuring type, currency, and depth). Depth refers to the number of distinct
contributors to each daily CDS spread collected by Markit. We use depth as a measure of CDS
liquidity (e.g., high depth indicates high liquidity see Qiu and Yu, 2012 and Lee, Naranjo, and
Sirmans, 2014 among others).14 We also gather recovery rate information from Markit. We match
each firms CDS to its corresponding sovereign CDS with the same contract characteristics, including
tier, restructuring type, and currency (see Internet Appendix A.2 for additional details).
3.1 Sovereign-level Variables
We collect country-level economic, institutional, and other control data from various sources. We
acquire each countrys sovereign S&P credit rating from Datastream and convert them into a nu-
merical score ranging from 0-21 (21 is equal to a AAA rating, 20 is equal to a AA+ rating, etc.)
to form a variable called Sovereign S&P Credit Rating. GDP and Government Debt-to-GDP are
downloaded from the International Monetary Fund (IMF). External government debt is retrieved
from the World Bank and scaled by GDP to form the variable External Debt-to-GDP. We collect
stock market index returns for each country from Datastream and compute a standard deviation of
weekly log returns each year to form local Stock Market Volatility.
To measure each countrys institutional and informational environments, we collect data from
several sources. We utilize the following four variables that measure the general strength of a
countrys property rights:
Property Rights measures the extent to which government creates and enforces laws that
protect private property and the extent to which government expropriates private property
(source: Heritage Foundation). 15
Rule of Law is an assessment of the law and order tradition of the country (source: Interna-
tional Country Risk Guide, ICRG).
14
Tang and Yan (2007) also use the total number of quotes and trades as a proxy for the level of CDS market
activity.
15
Dittmar and Yuan (2008) use Heritage Foundation data to measure each countrys economic freedom.
Repudiation Risk measures the countrys risk of contract repudiation or postponement due to
budgetary issues, political pressure, or a change in government (source: ICRG).
Expropriation Risk measures the extent to which a countrys government confiscates or na-
tionalizes private property or enterprise (source: ICRG).
Next, we use two variables that represent a countrys willingness to provide power and protection
to creditors:
Creditor Rights measures the legal protection allocated to creditors to have influence over
decisions that affect the value of their position. Specifically, this variable takes into account
(1) creditors voice in restructuring decisions, (2) automatic stay of creditors, (3) priority
given to creditors in bankruptcy, and (4) the degree to which creditors have control over the
firms assets during reorganization (source: La Porta, Lopez de Silanes, Shleifer, and Vishny,
1998; Djankov, McLiesh, and Shleifer, 2007).
Ln(Contract Enforcement Days) records the average number of days it takes to resolve a
creditor payment dispute in the courts. We use this variable as a measure of the efficiency
and enforcement of the legal system as it pertains to creditor rights (source: Djankov, McLiesh,
and Shleifer, 2007).
To measure a countrys level of informational transparency, we utilize two disclosure requirement

variables that capture distinct dimensions of disclosure. 16
Disclosure Requirements: Number of Items Reported records a countrys required number of

disclosed items on a wide range of topics, including general information, financial condition,
and corporate governance (source: Bushman, Piotroski, and Smith, 2004).
Disclosure Requirements: Reporting Frequency measures a countrys required timeliness of

financial reporting through the frequency of interim reports (source: Bushman, Piotroski,
and Smith, 2004).
16
Bushman, Piotroski, and Smith (2004) also use these firm-specific information environment variables as proxies
for country-level informational transparency.
All these institutional and informational variables are standardized to have a mean of zero and
standard deviation of one.
[ Insert Table 1 here ]
The country-level variable summary statistics are provided in Panel A of Table 1. There are 54
countries, with an average annualized 5-year sovereign CDS daily spread of 122.33 bps. The average
depth of the 5-year sovereign contract is 7.428. The average Ln(GDP) is 6.013 and is equivalent
to $408.70 billion (= e6.013 ). The average government debt issuance (Government Debt ) represents
55.101% of the countrys GDP.
3.2 Firm-level Variables
We gather firm-level financial data from Datastream, Worldscope, and Thomson One. From the
Thomson One database, we collect an annual S&P credit rating for each firm and convert it into a
numerical score ranging from 0-21 (21 is equal to a AAA rating, 20 is equal to a AA+ rating,
etc.). A difference of 1 in the firm rating score variable indicates a one sub-notch (+/-) difference.
We collect market capitalization, total assets, total debt, long-term debt, short-term debt, and
net cash flow at the end of each calendar year from Worldscope. The variable Size is defined as
the natural logarithm of market capitalization. Leverage is defined as total debt over total assets.
Short-term Debt/Total Debt is equal to short-term debt over the sum of short-term debt and long-
term debt. Cash Flow/Total Assets is defined as net cash flow over total assets. We also collect
each firms stock returns from Datastream and create Stock Return Volatility, which is computed
as the standard deviation of the firms weekly log returns in a given year. We compute Number of
Stock Exchanges by counting in Datastream the number of exchanges on which the firm has listed
its equity. For each firm-year, we calculate the expected default frequency ( EDFM erton ) of the
KMV-Merton default forecasting model according to Bharath and Shumway (2008) and Campbell,
Hilscher, and Szilagyi (2008) using weekly stock return volatility over the course of the year.
For our recovery rate analysis, we gather some additional variables. Profit Margin is computed
as EBITDA over Sales. Asset Tangibility is defined as Property Plant and Equipment over Total
Assets. Following Acharya, Bharath, and Srinivasan (2007), we compute several industry-specific
variables by country. Industry Asset Specificity is the industry median of the ratio of book value of
machinery and equipment to book value of assets. Industry Q is defined as the industry median of
the ratio of the firms market value (book value of assets book value of equity + market value of
the equity) to the book value of the firm (book value of assets). Industry distress dummy is equal to
one when the median stock return of firms within the industry is less than -30% and zero otherwise.
3.3 Scaled Net Exposure and Extra Disclosure Variables
Worldscope provides a variable on the percentage of total foreign assets, and it also provides the
amounts of total assets and total international assets so that foreign assets percentage can be
computed by dividing total international assets by total assets. We use the former to measure the
total foreign asset percentage, Foreign Assets/Total Assets. However, in instances where no foreign
assets percentage is explicitly provided, we use the computed value from the latter. We similarly
construct our total foreign sales percentage variable, Foreign Sales/Total Sales.
In addition to the aggregate view of foreign assets, Worldscope also provides geographically
segmented financial information. Up to ten geographic segments are reported, and companies
can input the description of the geographic segment as they choose. Thus, a firm has flexibility
to assign a single country or multiple countries to a single geographic segment. We count the
number of geographic segments reported by the firm to construct the variable Number of Geographic
Segments.17
To determine if a firms international asset exposure at the country-level has a net positive or
negative effect relative to the firms home country, it is necessary to compare the institutional char-
acteristics of the foreign countries in which the firm has assets with the institutional characteristics
of the firms home country. We address this issue by creating firm-level scaled net exposures
to various institutional and economic environment variables. We measure these variables net of
the home countrys value and scale the net exposure by Foreign Assets/Total Assets. Internet Ap-
pendix A.4 provides detailed steps on our construction of the scaled net exposure variables. Using
the foreign institutional characteristics associated with each of the firms asset segments, we define
We thank the Consulting Services Group at Thomson Reuters for providing missing and incomplete Worldscope
17
market segment data.

the Scaled Net Exposure variable (ScaledN etExposurei,t ) for firm-i in year-t as
ScaledN etExposurei,t =F orAsset%i,t

" 10 ! #
X SegAsseti,s,t (1)
SegInstV aluei,s,t HomeV aluei,t ,
s=1
F orAsset i,t
where F orAsset%i,t is Foreign Assets/Total Assets, SegInstV aluei,s,t is the institutional value of
segment-s, SegAsseti,s,t is the amount of assets in segment-s, F orAsseti,t is the total amount of
firm foreign assets, and HomeV aluei,t is the institutional value of the firms home country.
We create a firm-level Scaled Net Exposure variable using each of the following country-level
macroeconomic and institutional variables: GDP per Capita, local Stock Market Volatility, Prop-
erty Rights, Rule of Law, Repudiation Risk, Expropriation Risk, Creditor Rights, and Ln(Contract
Enforcement Days).18
A firm based in a country with low disclosure requirements may have to increase its informational
transparency when it lists its equity on a countrys exchange that has higher disclosure requirements.
To assess a firms level of transparency, we examine the disclosure requirements of the countries in
which the firm has listed its equity on an exchange. We create two Extra Disclosure variables for
firm-i in year-t, which are equal to the maximum disclosure index value among all countries that
list the firms equity in excess to the disclosure index value of the firms home country:
ExtraDisclosurei,t =
(2)
max [HomeExchangec , Exchangec1 , Exchangec2 , . . . ] HomeExchangec ,
where HomeExchangec refers to the disclosure requirement index value of the firms home country
and Exchangec1 , Exchangec2 , etc. refers to the disclosure requirement index values of each country
in which the firm lists its equity. Using this method, we construct Extra Disclosure: Number of
Items Reported and Extra Disclosure: Reporting Frequency.19
The firm level Scaled Net Exposure and Extra Disclosure variables are all standardized to
18
For example, Scaled Net Exposure: Property Rights measures the net positive or negative exposure by comparing
the Property Rights of the foreign countries in which the firm owns assets with the Property Rights of the firms
home country and then scales the exposure according to the total percentage of the firms foreign assets.
19
For example, if a firm based in Brazil has its equity listed in both Brazil and the United States, its overall level of
transparency as measured by Extra Disclosure: Number of Items Reported will be equal to the disclosure requirement
index value of the United States minus that of the Brazil.
have a mean of zero and standard deviation of one.

In Panel B of Table 1, we provide summary statistics on our firm-level variables. There are
2,364 firms from 54 countries. The average annualized 5-year firm CDS daily spread is 195.511 bps,
and the firm CDS contracts have an average market depth of 5.628. The average firm CDS recovery
rate from Markit is 37.042% with the standard deviation of 5.204%. 20 The average firm size in our
sample is $18.37 billion (= e23.634 ), with an average firm leverage of 29.9%. On average, firms in our
sample cross-list their stock on 2.348 different exchanges (Number of Stock Exchanges ) and operate
in the 3.693 different segments (Number of Geographic Segments ). To conserve space, Internet
Appendix A.1 and A.3 provides further details on our firm/country coverage and descriptions of
all our variables that are defined in this data section, respectively. The correlation matrix among
these variables is also provided in Internet Appendix A.6.
4 Main Results
4.1 Sovereign Risk: Local Institutional and Information Qualities
A central assumption of our main hypotheses developed in Section 2 is that investors price a
local governments T&C risk, and their T&C risk concerns are exacerbated as the institutional
and informational qualities of the local government deteriorate. T&C risk concerns could also
apply to government securities since both the enforceability of the sovereign financial contracts
and availability of information associated with those sovereign-issued claims differ across countries,
depending on their local institutional and informational qualities. We examine this assumption
using sovereign CDS contracts.
In Figure 3, we provide scatter-plots of each countrys natural logarithm of annual average 5-
year sovereign CDS daily spreads (Ln(Sovereign CDS Spread )) against each of our institutional
and informational factors using our full 54-country sample. As institutional factors, we consider
our four proxies for property rights protection (Property Rights, Rule of Law, Repudiation Risk, and
Expropriation Risk ) and two proxies for creditor rights protection (Creditor Rights and Ln(Contract
20
Based on our consultation with Markit, the recovery rate composite is the average of the passing recovery rates
submitted by their contributors. The contributors use these recovery rates as pricing inputs for their CDS trades.
Internet Appendix A.7 provides comprehensive descriptive statistics on Markits recovery rate data.
Enforcement Days)). For informational factors, we use Disclosure Requirements: Number of Items
Reported and Disclosure Requirements: Reporting Frequency.
We find strong graphical support for our key assumption. In the top four panels of Figure 3,
one can see that sovereign CDS spreads decrease in each of our four local property rights protection
measures (Property Rights, Rule of Law, Repudiation Risk, and Expropriation Risk ). In the following
two panels, a similar tendency is observed with our two local creditor rights protection proxies. This
tendency is clearly depicted in the panel where we use Ln(Contract Enforcement Days) as a proxy
for the local creditor rights. The plot shows that the longer it takes for creditors to seize their
recovery value from a defaulted institution, the higher is the institutions CDS spread. In the
last two panels of Figure 3, one can see a decreasing sovereign spread as the countrys local stock
exchange mandates stricter disclosure requirement in both the number of items to be reported and
the reporting frequency.
We further confirm the depicted inverse relation between sovereign CDS spreads and the local
institutional and informational qualities by running the following multivariate regression for country-
i, region-j, CDS contractual terms-k, and year-t:
Ln(Sovereign CDS Spread )i,j,k,t = 0 + k + t

+ Institutional (Informational) Factors i,t (3)
+ Ln(Region Sovereign CDS Spread )j,t + 0 Xi,t + i,t ,
where 0 is an overall constant, k is the vector of CDS contractual term (i.e., restructuring clause
and settlement currency) fixed effects, and t is the vector of year fixed effects. 21 Ln(Region
Sovereign CDS Spread) j,t is the natural logarithm of average CDS spreads for other countries in the
same region-j where country-i is located (j= North America, Latin America, Asia, Europe, and the
Middle East/Other ), and Xi,t is the vector of local market controls, including Stock Market Volatility,
Ln(GDP), Government Debt-to-GDP, and External Debt-to-GDP, all of which are well motivated
in the sovereign CDS and bond pricing literature (Longstaff, Pan, Pedersen, and Singleton, 2011;
Ang and Longstaff, 2013; Claessens, Klingebiel, and Schmukler, 2007; Ejsing and Lemke, 2011;
Dieckmann and Plank, 2012). We further control for Sovereign CDS Depth as a liquidity proxy and
cluster the standard errors at each country level.
21
All CDS contracts used in this paper have the same tier, senior unsecured tier, which is denoted as SNRFOR in
the Markit database.
The regression results are reported in Table 2. Consistent with the graphical evidence in Figure
3, we find significant local institutional and informational factor effects on sovereign CDS spreads.
For example, controlling for other factors, there is a 65.4% (i.e., 80 bps) decrease in the sovereign
CDS spread from its sample average (122.33 bps) for a one standard deviation increase in Property
Rights (column 1).22 Similar significantly negative effects are found for stronger creditor rights
protection (columns 5 and 6) and stricter information disclosure requirements (columns 7 and 8).
Overall, the results in Table 2 provide evidence consistent with our key assumption that investors
T&C risk concerns are closely related to the local governments institutional and informational
qualities. We show that such risks are priced in ways we assumed at the sovereign CDS contract
level.23
4.2 Exodus from Sovereign Risk: Global Asset Geographic Location

and Information Networks
How can a firm opt out of its local legal and institutional heritages? In this section, we test whether
a firms global asset and information networks determine its effective institutional and information
environments, thereby affecting the firms credit pricing.
4.2.1 The Determinants of Corporate CDS Spreads
We first conduct a univariate analysis to examine the extent to which a firms simple multinational
status affects its CDS spread. In Panel A of Table 3, we use one of the following variables as proxies
for a firms multinational status: 1) No. of Stock Exchanges, 2) No. of Geog. Segments, 3) Foreign
Assets/Total Assets, and 4) Foreign Sales/Total Sales. Using each variable, we sort our sample
firms and put them into four different buckets: Zero, Low, Medium, and High. Here Zero implies
zero international exposure. As sovereign risks became more serious concerns during the recent
crisis period, we further decompose our sample into two sub-periods: pre-crisis period (2004-2007)
and crisis period (2008-2011). Then, we compute the average firm CDS spread in each bucket for
22
All local Institutional (Informational) Factors are normalized to have a mean of zero and standard deviation of
one for interpretation and inference ease.
23
In each model specification in Table 2, our control variables typically have the correct expected signs and
statistical significance (positive sign for Ln(Region Sovereign CDS Spread), Stock Market Volatility, Government
Debt (% of GDP), and External Debt (% of GDP); negative sign for Ln(GDP)). However, our liquidity measure,
Sovereign CDS Depth, has the opposite sign to our prior.
the two sub-periods and report the results in Panel A of Table 3.

In Panel A of Table 3, one can see that average corporate CDS spreads in both sub-periods
are lower for the firms whose stocks are cross-listed in multiple foreign exchanges ( No. of Stock
Exchanges) and whose assets are located in multiple countries (No. of Geog. Segments ). When
we sort the firms based on each firms aggregate foreign assets (Foreign Assets/Total Assets ) and
sales amount (Foreign Sales/Total Sales), we find similar patterns. In all cases, the CDS spreads
of multinational firms (High) are lower than purely domestic companies (Zero), and the differences
increase in their magnitudes during the crisis period. For example, based on Foreign Assets/Total
Assets, there is a -22.549 bps difference between the two types of firms (High - Zero) in the pre-crisis
period, whereas a -113.558 bps difference between them during the crisis.
Arguably, the results in Panel A could simply reflect that MNCs are better quality firms than
purely domestic entities. To provide more direct evidence that foreign asset geographic locations
matter for corporate CDS spreads, in Panel B of Table 3, we incorporate the locational information
of each MNCs foreign assets and construct the effective institutional qualities that the multinational
firm is exposed to across its various markets. We use our scaled net institutional exposure measures
developed in Section 3.3 as sorting variables. We consider our four property rights proxies and two
creditor rights proxies to show the relevance of these scaled net institutional exposures on a firms
average CDS spread.
The results in Panel B show that CDS spreads of MNCs whose foreign assets are exposed to
better property rights institutions (Net Positive Exposure) have lower average CDS spreads than
those exposed to poor property rights institutions (Net Negative Exposure). We also find a similar
tendency with creditor rights protection measures. The difference in average corporate CDS spreads
between Net Positive Exposure and Net Negative Exposure firms are widened during the crisis period,
suggesting that our proposed delinking channels from sovereign risk become more effective during
the time when investors T&C risk concerns are more severe.
To further control for other factors that influence corporate CDS spreads, we specify our re-
gressions taking into account the formal corporate credit risk models explored by Collin-Dufresne,
Goldstein, and Martin (2001), Bharath and Shumway (2008), and Campbell, Hilscher, and Szi-
lagyi (2008). We test for the effects of each firms global asset and information networks on their
CDS spreads running the following annual panel regression for firm- i, industry-j, country-k, CDS
contractual terms-l, and year-t:
Ln(Firm CDS Spread )i,j,k,l,t = 0 + t + j + k + l

+ ScaledN etExposure(ExtraDisclosure)i,t (4)
+ 0 Yk,t + 0 Xi,t + i,t ,
where Ln(Firm CDS Spread ) is the natural logarithm of annual average 5-year corporate CDS
daily spreads (in bps). On the right-hand-side of the regression specification, we have 0 , the
overall constant, t , the vector of year fixed effects, and j , k , and l , the vectors of industry,
country, and CDS contractual term fixed effects, respectively. 24
ScaledN etExposure(ExtraDisclosure)i,t are the main explanatory variables that fully incor-
porate the locational information on a firms foreign assets and its equity cross-listing on foreign
exchanges. We use: 1) Scaled Net Exposure: Property Rights for property rights protection; 2)
Scaled Net Exposure: Creditor Rights Protection for creditor rights protection; and 3) Extra Disclo-
sure: Number of Items Reported for disclosure requirement. 25 All these variables are normalized to
have a mean of zero and standard deviation of one to facilitate economic magnitude interpretations.
Yk,t is the vector of sovereign-level time-varying risk and characteristic controls, including Ln(Sovereign
CDS Spread), Ln(GDP per Capita), Government Debt-to-GDP, and Stock Market Volatility. Xi,t ,
the vector of firm-level controls, including Size, Leverage, Short-term Debt/Total Debt, Cash Flow/Total
Assets, Excess Stock Return, and Stock Return Volatility. The firm-level controls also include an
ex ante structural measure of default risk, EDFM erton , which we estimate by solving a system of
two nonlinear equations from the Merton (1974) model following Bharath and Shumway (2008)
and Campbell, Hilscher, and Szilagyi (2008). We also control for potential liquidity effects using
corporate CDS contract Depth. We cluster standard errors at the firm level to adjust for potential
within-firm regression residual persistence.
24
The inclusion of year fixed effects alleviates potential concerns of any omitted time-varying global factors that
may affect both the firm and sovereign CDS spreads (Longstaff, Pan, Pedersen, and Singleton, 2011). Country
fixed effects implicitly control for any regional fixed effects. Industry heterogeneity in corporate credit spreads is
further captured by the industry fixed effects. Each firms CDS contract is matched with its local governments CDS
contract based on their restructuring clauses and base currencies. For potential contractual term variation across
these matched firm-sovereign CDS pairs, we control for restructuring type and settlement currency dummies.
25
We provide additional regression result robustness tests using other proxies for property rights, creditor rights,
and information disclosure requirements in Internet Appendix B.1.
Panel A of Table 4 reports our main results. In columns 1 to 3, one can respectively see the
significantly negative effects of Scaled Net Exposure: Property Rights, Creditor Rights, and Extra
Disclosure: No. of Items Reported on the natural logarithm of annual average 5-year corporate
CDS spreads. In column 1, the point estimate on Scaled Net Exposure: Property Rights is -0.0519,
indicating that for a one standard deviation increase in that variable, there is approximately a
10.147 bps (5.19%) decrease in corporate CDS spread from its sample average (195.511 bps). This
effect is statistically significant at the 1% level. In columns 2 and 3, we respectively see 14.292 bps
(7.31%) and 20.333 bps (10.4%) decreases in corporate CDS spreads as the other two variables,
Scaled Net Exposure: Creditor Rights and Extra Disclosure: No. of Items Reported, increase by
one standard deviation in their values. In column 4, where we jointly use all of our scaled net
institutional exposure and extra disclosure variables, we find that all of them significantly explain
Ln(Firm CDS Spread). These factors jointly explain a 40.393 bps (20.66%) decrease in Ln(Firm
CDS Spread) for a one standard deviation increase in their values.
In column 5 of Table 4, we further show the robustness of our results to outliers by running a
quantile regression. In this regression, we use the annual average of 5-year corporate CDS daily
spreads as our dependent variable. As independent variables, we use the 5-year sovereign CDS
spread level (Sovereign CDS Spread ), instead of Ln(Sovereign CDS Spread), as one of the sovereign
risk controls, while all of the other control variables are the same as those used in our previous linear
regression specifications. Standard errors are constructed using a block bootstrapping method with
300 replications, where the bootstrapping block is defined at the firm level. In column 5 we find an
average 17.806 bps (= -5.884 -6.598 -5.324) combined effect of our scaled net institutional exposure
and extra disclosure variables on the 5-year corporate CDS spreads. These effects are statistically
significant at the 1% level.
The point estimates of the firm- and sovereign-level controls in Panel A of Table 4 have their
expected signs: positive signs for Ln(Sovereign CDS Spread) (or Sovereign CDS Spread ), Leverage,
Short-term Debt/Total Debt, Stock Return Volatility, and EDFM erton ; negative signs for Size, Cash
Flow/Total Assets, and CDS Depth. One exception is the positive point estimate on Excess Stock
Return, though it is statistically insignificant in all five columns of Panel A.
While the results in Panel A of Table 4 show that firms can reduce their CDS spreads through
their global network connections to better institutional and informational environments, the re-
sults do not necessarily imply that firms can strictly decouple themselves from their sovereign risks
through these channels. To directly test that our institutional and informational channels are ef-
fective mechanisms for firms to exodus their sovereign risks, we examine sovereign ceiling rule
violations. In these auxiliary tests, we use alternative Probit and Ordered-Probit model specifi-
cations to examine whether our factors explain various SCV indicators (binary and multinomial
indicators) based on the corporate-sovereign CDS spread differences. We find that both our insti-
tutional and informational factors significantly explain the intensity of a CDS market SCV during
the crisis period. We discuss these results in Internet Appendix D.
4.2.1.1 Locational Information, Scaled Net Foreign Fundamentals, and Crisis Period
Interaction Effects
In Panel B of Table 4, we provide some additional identification checks. Each specification in

Panel B is similar to those we ran in Panel A of Table 4, but in Panel B we include additional
control variable tests. The first half of Panel B (columns 1-5) shows robustness checks to aggregate
international exposure and scaled net foreign fundamentals, while the second half of the panel
(columns 6-9) provides crisis period interaction effects.
In columns 1-3, we sharpen our identification on the locational information of a firms foreign
asset holdings and its stock cross-listing status by including additional controls that capture each
firms aggregate foreign exposure: Foreign Assets/Total Assets, Foreign Sales/Total Sales, and
Number of Stock Exchanges. One can see that the point estimates of our Scaled Net Exposure:
Property Rights, Creditor Rights, and Extra Disclosure: No. of Items Reported variables are little
affected by controlling for each of the three additional aggregate foreign exposure measures. These
results confirm that it is the geographic segmentation information of a firms assets, not the aggre-
gate foreign exposure itself, that explains the firms CDS spread (columns 1 and 2). Moreover, a
simple measure of a firms foreign stock exchange exposure (Number of Stock Exchanges ) does not
fully explain the effect of our informational channel (column 3). The requirement by foreign stock
exchanges on a firms extra information disclosure explains a 4.83% reduction in corporate CDS
spread from its sample average even after controlling for the Number of Stock Exchanges variable.
Another potential concern is that our scaled institutional and informational variables could
simply capture a firms exposure to foreign fundamentals, such as a foreign countrys GDP (per
Capita) and that countrys Stock Market Volatility, not the countrys institutional and informational
qualities. To address this possibility, we additionally control for scaled foreign country fundamental
exposures, Scaled Net Foreign Fundamentals: Ln(GDP per Capita) and Stock Market Volatility, in
columns 4 and 5 of Panel B of Table 4. We again find that our institutional and informational
channels are robust to the foreign fundamental exposure concerns.
T&C risk concerns are expected to be more severe to (foreign) investors when the local gov-
ernment is in distress, with a high default probability. From Figure 1, we show that there is a
dramatic increase in global 5-year sovereign CDS spreads, and we therefore expect that the effects
of our Scaled Net Exposure: Property Rights, Creditor Rights, and Extra Disclosure: No. of Items
Reported variables are stronger during this time period. These additional tests also enable us to
further identify the T&C risk channel through which our scaled net institutional and informational
factors affect corporate CDS spreads. To this end, we interact each of our three scaled net exposure
factors with a crisis indicator variable that takes a value of one for 2008-2011 and zero otherwise.
In all four columns 6 to 9 in Panel B of Table 4, we find significant Scaled Net Exposure: Property
Rights and Creditor Rights and Extra Disclosure: No. of Items Reported effects in explaining the
annual average corporate CDS spread, especially during the crisis period. For example, in column 9,
the coefficient estimate on the interaction term between Scaled Net Exposure: Property Rights and
the crisis dummy (-0.0417) is approximately 3.53-times larger than its stand alone effect (-0.0118).
The interaction effect is statistically significant at the 5% level. We also find similar significance
patterns with the other two variables, Scaled Net Exposure: Creditor Rights and Extra Disclosure:
No. of Items Reported.
4.2.1.2 Do Corporate Credit Ratings Incorporate Global Asset and Information Net-
work Information?
To what extent do ratings agencies account for geographical foreign asset segmentation information
or foreign equity cross-listing disclosure requirement information in their corporate credit rating
assignments? Do they simply account for the firms aggregate foreign operations or fully reflect the
firms global asset and information networks? More importantly, does the international CDS market
provide richer information content than major rating agencies? In this section we address these
questions by using an experimental design similar to our earlier corporate CDS spread analysis.
Here we use the S&Ps corporate credit rating score as a main dependent variable. 26
Table 5 provides our S&P credit rating analysis regression results. In columns 1 and 2, we first
show that S&P corporate credit ratings account for aggregate foreign exposure, Foreign Assets/Total
Assets and Foreign Sales/Total Sales, information. Both variables are positively associated (0.974
and 1.415 for Foreign Assets/Total Assets and Foreign Sales/Total Sales, respectively) with Firm
S&P Credit Rating, which takes a higher numeric value for a better credit quality (AAA=21,
AA+=20, ...). However, in columns 3 to 5 where we instead use each of our scaled net institutional
exposure and extra disclosure variables one at a time, we find that except Extra Disclosure: No. of
Items Reported, none of the other two Scaled Net Exposure: Property Rights and Creditor Rights
variables is reflected in the Firm S&P Credit Rating. We get similar results in column 6 where we
jointly include the three measures in the regression. Overall, the results in columns 1 to 6 of Table 5
suggest that S&Ps corporate credit ratings do not substantively reflect a firms geographical foreign
asset information, but do account for a firms cross-listing location information in its corporate rating
assignments.
To further show that CDS market investors more efficiently incorporate a firms foreign asset
geographical information into their corporate CDS pricing, we re-run our earlier 5-year corporate
CDS spread regression while controlling for each firms Firm S&P Credit Rating in place of the firm
fundamentals. Looking at the results in column 7 of Table 5, we find that corporate CDS spreads
reflect a firms global asset geographical exposure information that is not captured by the S&P
credit ratings. Our two net institutional quality proxies, Scaled Net Exposure: Property Rights and
Creditor Rights, significantly explain a 7.48% (=0.0351+0.0397) reduction in 5-year corporate CDS
spreads above and beyond the effect associated with S&P credit ratings. 27
26
In untabulated regression results, we also use composite credit ratings provided by Markit, which are based on
the ratings from all major rating agencies. These composite ratings do not have sub-notch information. However, we
get both quantitatively and qualitatively similar results to those using S&P credit ratings with sub-notch information.
Note that the sample size for the ratings analysis is smaller because of unrated entities. Our result are also robust
to using log transformed corporate credit ratings as an alternative dependent variable. These two sets of results are
available in Internet Appendix B.3 and B.4, respectively.
27
These results suggest that S&P does not actively incorporate firm-level global asset network connections in their
corporate credit rating assignments. Related to this point, we further discuss in Internet Appendix D a potential
information spillover on SCVs from CDS markets to S&P credit ratings. We provide detailed evidence that SCV
information spills over from 5-year CDS spreads to S&P credit ratings; for example, we find that the 5-year CDS
SCVs precede the S&P credit rating SCVs by two to three years. However, we do not find that the opposite holds.
4.2.2 The Effects of the Expected Recovery in the Event of Default: Institutional
versus Informational Channels
As a further identification test, we also examine the effects of our scaled net institutional and
informational factors on the expected recovery rates of corporate CDS contracts in the event of
default. As hypothesized in Section 2, we expect that our institutional channel better explains the
overall efficiency of a post-default restructuring process than our informational channel because
the rules and conditions governing a firms assets following default tend to be more closely tied to
the degree of property rights and creditor rights protections. Although we expect that better firm
financial information availability reduces investors forecasting errors on expected recovery rates in
the event of default, the recovery level is arguably more likely to be explained by the strength of
legal and institutional conditions.
In our recovery rate tests, we use Markits expected recovery rates on corporate CDS contracts.
Based on our consultation with Markit, their recovery rate composite is the average of the pass-
ing recovery rates submitted by their contributors. The contributors use these recovery rates as
pricing inputs for their CDS contributed spreads, and these recovery rate contributions are subject
to Markits same data cleaning tests as the contributed CDS spreads. 28 Using the natural loga-
rithm of the corporate CDS expected recovery rate composite (Ln(Firm CDS Recovery Rate)) for
firm-i, industry-j, country-k, CDS contractual terms-l, and year-t, we run the following regression
specification explored by Acharya, Bharath, and Srinivasan (2007):
Ln(Firm CDS Recovery Rate)i,j,k,l,t = 0 + t + j + k + l

+ ScaledN etExposure(ExtraDisclosure)i,t (5)
+ 0 Yk,t + 0 Zj,t + 0 Xi,t + i,t ,
28
Internet Appendix A.7 provides comprehensive descriptive statistics on Markits recovery rate data. Internet
Appendix B.5 and B.6 provide additional robustness checks to our results reported in Table 6 of this main text. We
find substantial temporal and cross-sectional variation in the composite recovery rates across ratings, restructuring
clauses, currencies, and the reference entities countries of domicile. In Internet Appendix A.7 we also show that
Markit recovery rates foreshadow and converge towards realized recovery rates for a sample of defaulted firms that we
retrieved from Creditex (website: www.creditfixings.com). There are several other studies that also use the Markit
recovery rate composite, including Berndt and Obreja (2013) and Berndt (2014). Because of our extensive sample of
firms across countries and the need for time varying recovery rates in our analysis, it is difficult to comprehensively
estimate risk-neutral recovery rates from parametric models using the term structure of CDS spreads ( Pan and
Singleton, 2008; Elkamhi, Jacobs, and Pan, 2012), equity/equity option data (Jarrow, 2001; Carr and Wu, 2009),
among other approaches.
where ScaledN etExposure(ExtraDisclosure)i,t are the main explanatory variables, and all s are
the same fixed effect vectors as in our corporate CDS spread regression (Equation 4). Yk,t is the
vector of sovereign-level time-varying risk and characteristic controls and is similar to those used
in Equation (4), except we replace Ln(Sovereign CDS Spread) with Ln(Sovereign CDS Recovery
Rate).29 As the vector of firm-level controls, Xi,t includes Size, Leverage, Profit Margin, and Asset
Tangibility.
Following Acharya, Bharath, and Srinivasan (2007), we also introduce the following vector of
industry-level control variables (Zj,t ): Industry Asset Specificity, Industry Q, and Industry Distress
Dummy. Standard errors are clustered at each firm level.
The results in Table 6 support our hypotheses; the institutional channel matters more for cor-
porate CDS contract expected recovery rates than the informational channel. In column 4 of Table
6, where we jointly use our scaled net institutional and informational factors, we find that Scaled
Net Exposure: Property Rights and Creditor Rights significantly explain Ln(Firm Recovery Rate)
at least at the 5% level. Their economic significance is also high, explaining a 3.16% (= 0.0199
+ 0.0117) combined effect on the corporate CDS recovery rate from its sample average (37.042%).
However, as expected, we do not find that our Extra Disclosure: No. of Items Reported variable
significantly affects corporate CDS recovery rates. Overall, the recovery rate results in Table 6 are
consistent with our hypotheses, rendering further support for our underlying T&C risk channels. 30
4.2.3 Selection and Reverse Causality: Causal Evidence
While we have shown that our institutional and informational channels are well identified to the
inclusion of additional controls and also extend to recovery rate tests, there are still some potential
selection and reverse causality concerns. Arguably, an alternative interpretation of our results could
potentially be that instead of our regressions in Equation (4) capturing the effects of better laws,
institutions, and information, our scaled net institutional and informational factor effects could
simply be capturing better quality multinational firms that can place their assets and/or cross-list
29
The expected recovery rates on sovereign CDS contracts are also collected from Markit.
30
The point estimates of our control variables are consistent with their expected signs and the findings in Acharya,
Bharath, and Srinivasan (2007): Positive signs for the point estimates on Size, Profit Margin, Asset Tangibility,
Industry Q, whereas negative signs on Leverage, Industry Asset Specificity, and Industry Distress Dummy. We also
expect the positive point estimate sign on Ln(Sovereign CDS Recovery Rate).
their stocks in better institutions than their home market as they aim to reduce their financing
costs.
To structurally illustrate these strong firm selection biases, consider Equation ( 4) represented
in the following short form:
Si,t = + Fi,t + i,t , (40 )
where Si,t and Fi,t respectively denote firm-is natural logarithm of annual average 5-year CDS
spreads and its scaled net institutional (or extra disclosure) factor in year- t. denotes a vector of
various constant terms.
Suppose that there is an omitted firm quality variable (Qi,t ) such that i,t = Qi,t + i,t , where
for F becomes:
< 0 and E[F 0 ] = 0. Then, our estimated coefficient, ,
= Cov(S,F )
V ar(F )
Cov(+F +Q+,F )
= V ar(F )
(6)
= + Cov(Q,F
V ar(F )
)
Given < 0, if Cov(Q, F ) > 0, then could be spuriously negative although the true effects of our
factors do not exist (i.e., = 0). A key observation here is that the strong firm effect exists only
if Cov(Q, F ) > 0 (i.e., better (weaker ) quality firms than purely domestic entities can place their
assets and/or cross-list their stocks in better (weaker ) institutions than their home institutions).
To further see the source of this key correlation condition, recall that our scaled net institutional
factor, F , takes the following short form:
X
Fi,t = Ps,t i,s,t , (10 )
sSeti,t
where Seti,t is the set of foreign segments of firm-i in year-t, Ps,t is the net institutional quality
of segment-s above and beyond the home countrys value, and i,s,t is firm-is foreign asset weight
in segment-s in year-t. Ps,t has little time variation, and therefore the strong firm selection
biases center at {i,s,t |s Seti,t } the endogenous distribution of a firms foreign asset geographic
locations. The endogenous asset distribution should satisfy the following conditions to ensure
Cov(Q, F ) > 0:
i,s,t
Qi,t
> 0 for Ps,t > 0
(7)
i,s,t
Qi,t
< 0 for Ps,t < 0.
For our extra disclosure factors, a similar concern lies at the endogenous distribution of the firms
foreign stock exchanges where its stock is cross-listed.
Based on our structural layout above on the sources of the strong firm effects, we tackle this
potential strong firm issue in the following three ways: (1) For a constant unobserved firm quality
(Q = Qi ), we directly control for firm fixed effects. For a time-varying firm quality ( Q = Qi,t ),
we either (2) identify a firms foreign market exposure that satisfies Cov(Q, F ) 0 or (3) directly
introduce foreign market institutional quality shocks (i.e., shocks in Ps,t ) where the firm has its
asset holdings but cannot ex-ante correctly anticipate (or circumvent) the shocks. Using these two
conditions, we construct instruments for F and rule out the strong firm effect argument. We
report these results in Table 7.
In columns 1-4 of Panel A in Table 7, we report the results using our first approach the firm
fixed effect controls. In all four columns, we find statistically significant CDS spread reductions
associated with all of our scaled net institutional and extra disclosure exposures. These effects are
also economically significant, explaining a 22.191 bps reduction (-11.35%= -0.0244 -0.0427 -0.0464)
in 5-year corporate CDS spreads for a one standard deviation increase in their values (column 4). 31
In columns 5-7 in Panel A, we report the results using our second approach. It is important to
note that MNCs whose assets reside outside their home country tend to be stronger firms than
purely domestic entities regardless of their foreign asset geographic locations. 32 With this key obser-
vation in mind, one can see from the second condition in Equation (7) that any MNC foreign asset
exposure in weaker institutions than their home institutions (i.e., {i,s,t |s Seti,t s.t. Ps,t < 0})
can serve as an instrument to address selection biases. To the extent that not even strong MNCs
31
We find similar results, both economically and statistically, when we use the change-on-change regression speci-
fications. See Internet Appendix B.7 for more details of these results.
32
Internet Appendix B.9 provides direct evidence showing that MNCs going to weaker institutions than their home
institutions over our sample are indeed stronger firms than purely domestic entities over several important firm
characteristics. Anecdotal evidence also suggests that MNCs with strong brand names tend to go abroad to weaker
institutions than their home institution as their home market has lost its appeal as a place to invest and hire. See, for
example, Wall Street Journal : Big U.S. Firms Shift Hiring Abroad (David Wessel, April 19, 2011). The literature
also identifies several reasons for MNC foreign investment, including market seeking to use their unique or superior
products to compete in foreign markets and resource seeking to obtain less costly access to the production inputs
such as land, labor, capital, and natural resources (Dunning, 2002).
can overcome the net negative institutional and informational environment effects on their CDS
spreads, an instrumental variable (IV) estimation based on this net negative foreign market ex-
posure would suggest that it is the asset location placement in the institutional and informational
environment that increases these MNCs spreads, and that these net negative foreign exposure
results will capture the lower bound of our global asset and information network effects on corporate
borrowing costs.
The IV estimate becomes analogous to separately estimating the effects of net negative
P
(i.e., F = sSeti,t ,Ps,t <0 Ps,t i,s,t ) and net positive foreign exposure factors (i.e., F + =
P
sSeti,t ,Ps,t >0 Ps,t i,s,t ) on 5-year corporate CDS spreads, because the net negative foreign ex-
posures are strictly a subset of our original scaled net institutional and informational variables.
The results in columns 5 and 6 of Panel A confirm that the net negative foreign asset exposure
significantly increases 5-year corporate CDS spreads by 5.58% and 3.64% through Property Rights
and Creditor Rights channels relative to purely domestic entities. In column 7 of the same panel, we
show for our Extra Disclosure: No. of Items Reported variable that 5-year corporate CDS spreads
increase by 4.19% following delisting events (Delisted from any exchange), but they particularly do
so for net negative information exposure events with relaxed information disclosure requirements
(21.7%).33
In Panel B of Table 7, we show the results from using our third approach that is based on large
unexpected shocks in the degree of foreign country property rights protection where a respective
MNC has its asset holdings. We again focus on the large negative shocks in Property Rights
corresponding to less than 5% events in the left tail of the variables distribution. 34 We choose
negative shocks rather than positive shocks because we want to be conservative about the
possibility that MNCs could forecast even these rare events and preemptively adjust their asset
holdings before the large shocks hit the foreign countries. Given that MNCs would likely reduce
33
It should be noted that Extra Disclosure: No. of Items Reported is defined as the difference between the
maximal information disclosure requirement of any exchange where a firms stock is cross-listed and the requirement
mandated by its home stock exchange. By definition, this variable cannot be negative it is bounded below by
zero. We therefore estimate delisting event effects, focusing particularly on events where the extra information
disclosure requirements are relaxed following the delisting. To validate our instrument, we show that these events
are not associated with weak firm fundamentals. Internet Appendix B.10 shows that firm-years that underwent
these Delisting from high disclosure exchange events are in fact associated with better firm quality over several key
characteristics than those that did not delist and/or went through the Delisted from low disclosure exchange events.
34
Because Creditor Rights and Disclosure Requirements variables are time-invariant, we focus on annual events
with large negative property rights shocks in this analysis. The distributional summary of the changes in property
rights is provided in Internet Appendix B.8.
their foreign asset holdings before the large negative shocks hit, a foreign asset exposure effect in
these countries, if any, would capture the lower bound of the negative Property Right shock effects.
These events also occur in foreign countries, and are thus unlikely to be correlated with local factors
that affect both the treated firms and their local sovereign governments credit qualities. 35
In Figure 4, we plot the time series of the treated firms average asset holdings. This figure
provides evidence that these negative property rights shock events are ex ante difficult to predict
(or circumvent) by the treated firms. One can see that the treated firms actually increase their
asset holdings in those countries before the treatment year-0 and reduce their holdings only after
the treatment.
Despite these temporal asset holding patterns, one could still argue that these negative property
rights shocks in foreign countries are still potentially endogenous because sovereign fundamentals
tend to be regionally integrated (Ang and Longstaff, 2013; Longstaff, Pan, Pedersen, and Singleton,
2011) and even such rare events in foreign countries could be closely followed by treated firms
managers. To further increase both firm and sovereign informational and other potential linkage
barriers, we introduce the following additional filters on our foreign negative shocks based on the
literature:36 (1) geographically distant by more than 4,000 miles; (2) culturally distant by more
than a one standard deviation measured by the World Values Survey; (3) use different language
(also excludes global common languages English, French, German, and Spanish) from the treated
firms home countries; and (4) in consultation with the Heritage Foundation, narrow country tests
that had large and exceptional property rights related events, including nationalizations (Argentina
(2009), Venezuela (2008)), eliminations of democratic processes (Egypt (2007), Guinea (2009)), and
deadly crackdown on protesters by military ruler (Guinea (2009)), all of which are arguably more
exogenous or ex ante harder to circumvent. 37
35
If we use large negative shocks that directly affect the degree of property rights protection in the MNCs home
country, we cannot clearly identify the foreign asset geographic location effects on their CDS spreads. The effects
could be confounded with the effects of unobserved local fundamental shocks that deteriorate the firms credit quality.
36
An extensive literature documents the significant impact of distance, culture, and language on economic and
financial connectivity, exchanges, and outcomes, with proximity increasing the linkages ( Grinblatt and Keloharju,
2001; Guiso, Sapienza, and Zingales, 2009; Siegel, Licht, and Schwartz, 2011; Giannetti and Yafeh, 2012). We
use these variables as additional separating mechanisms to further identify the causal direction of our institutional
channel.
37
The worldwide transportation network literature indicates that up to 90% of the direct links in shipping traffic
occur at less than 3,700 miles and that the global cargo ship network accounts for 90% of the international exchange
of goods (Ducruet and Notteboom, 2012; Woolley-Meza, Thiemann, Grady, Lee, Seebens, Blasius, and Brockmann,
Using treatment dummies for each of these unexpected negative foreign property rights shocks,
we run two stage least squares instrumental variable (2SLS IV) regressions using the specification
in Equation (4). In column 1 of Panel B in Table 7, we use our full treatment sample without
any additional filter and find a statistically significant and negative point estimate on Scaled Net
Exposure: Property Rights (-1.006) in the second stage at the 1% level. In columns 2 and 3 of Panel
B in Table 7, where we use the one- or two-year lagged distributions of each treated firms asset
holdings, we find nearly identical (-1.010 in column 2 and -0.833 in column 3) average treatment
effects on corporate CDS spreads. In columns 4 to 7 of Panel B where we impose our additional
filters based on geographical distance (column 4), cultural distance (column 5), language barriers
(column 6), and finally narrow country events (column 7), we continue to confirm the negative and
statistically significant point estimates of Scaled Net Exposure: Property Rights in the second stage
regressions.
Overall, all our causal regression results in Table 7 suggest that it is the asset location placement
in the institutional and informational environment that affects the corporate CDS spreads and that
not even strong firms can avoid these negative institutional and informational environment effects.
To conserve space, we further show in Internet Appendix B.12 that our treatments in Panel B
of Table 7 are unlikely to be confounded with various firm-level observables by conducting nearest-
neighbor matching estimations of the average treatment effect on the treated using either Maha-
lanobis distances or propensity scores estimated from the Probit regression. 38 We also report results
using dynamic treatment dummies from two-year prior to one-year after the actual treatment year.
We find that all our treatment effects only exist after the actual treatments, not before, which
further confirms the sharp timing identification of our treatments.
5 Conclusion
Difficult economic and political situations in countries often result in governments taking actions
that deprive or restrict a companys ability to operate. In this paper, we explore the extent to which
2011). However, we also show that our results are robust to the use of an alternative distance cutoff, 5,000 miles.
Similarly, for the cultural distance, we show that our results are robust to the use of two standard deviations as
an alternative cutoff. All these results are available in Internet Appendix B.11. For in-depth information on events
occurring in our narrow country tests, see the Internet Appendix A.8.
38
For matching on Mahalanobis distance, we report Abadie and Imbens (2006, 2011) bias-corrected matching
estimator.
firms can escape these sovereign risks through their use of global asset and information networks.
We hypothesize that both the institutional and informational qualities of a local government are
closely related to the degree of investors concerns regarding the governments transfer and con-
vertibility risk, termed T&C risk, as well as more extreme exporiatory governmental actions. We
propose two novel channels (Institutional and Informational) through which private sector com-
panies could delink themselves from their local governments T&C type risks, thereby alleviating
investors concerns: 1) placing firm assets in foreign countries with better property and creditor
rights protection, and 2) cross-listing their stocks on foreign exchanges where stricter information
disclosure is mandated.
Using 5-year CDS spreads on 2,364 companies in 54 countries during 2004-2011, we provide
evidence showing that firms exposed to better property rights institutions through their foreign
asset positions and firms whose stocks are listed on exchanges with stricter disclosure requirements
reduce their CDS spreads by 40 bps for a one standard deviation increase in their exposure on the two
channels. These channels capture distinct effects beyond those associated with firm- and country-
level fundamentals, and the channels effects increase during the recent sovereign credit crisis as
governments increased their risk transfer actions to private sector entities. We also find that our
institutional channel better captures CDS recovery rate effects. We provide numerous robustness
checks, showing that our results are robust to potential specification, identification, selection, and
causality concerns. Overall, our results contribute towards three important research streams on
multinational corporate credit risk pricing, international corporate linkages and exposure, and legal
and institutional heritage effects.
Taken together, our results suggest that firms can, in part, avoid their sovereign governments
potential grabbing risk exposure through firm-level global asset and equity cross-listing connections
in countries with better institutional and informational environments.
References
Abadie, Alberto, and Guido W Imbens, 2006, Large sample properties of matching estimators for average treatment
effects, Econometrica 74, 235267.
, 2011, Bias-corrected matching estimators for average treatment effects, Journal of Business and Economic
Statistics 29, 111.
Acharya, Viral V, Sreedhar T Bharath, and Anand Srinivasan, 2007, Does industry-wide distress affect defaulted
firms? Evidence from creditor recoveries, Journal of Financial Economics 85, 787821.
Acharya, Viral V, and Timothy C Johnson, 2007, Insider trading in credit derivatives, Journal of Financial Economics
84, 110141.
Ang, Andrew, and Francis A Longstaff, 2013, Systematic sovereign credit risk: Lessons from the U.S. and Europe,
Journal of Monetary Economics 60, 493510.
Bae, Kee-Hong, and Vidhan K Goyal, 2009, Creditor rights, enforcement, and bank loans, The Journal of Finance
64, 823860.
Bai, Jennie, and Shang-Jin Wei, 2012, When is there a strong transfer risk from the sovereigns to the corporates?
Property rights gaps and CDS spreads, Georgetown University and Columbia University Working Paper.
Bailey, Warren, Andrew G Karolyi, and Carolina Salva, 2006, The economic consequences of increased disclosure:
Evidence from international cross-listings, Journal of Financial Economics 81, 175213.
Bekaert, Geert, Campbell Harvey, and Christian Lundblad, 2005, Does financial liberalization spur growth?, Journal
of Financial Economics 77, 355.
Bekaert, Geert, and Campbell R Harvey, 2000, Foreign speculators and emerging equity markets, The Journal of
Finance 55, 565613.
Bergman, Nittai K, and Daniel Nicolaievsky, 2007, Investor protection and the Coasian view, Journal of Financial
Economics 84, 738771.
Berndt, Antje, 2014, Decomposing CDS spreads and their variation, North Carolina State University Working Paper.
, Robert A Jarrow, and ChoongOh Kang, 2007, Restructuring risk in credit default swaps: An empirical
analysis, Stochastic Processes and their Applications 117, 17241749.
Berndt, Antje, and Iulian Obreja, 2013, Sources of common variation in changes in CDS spreads: Theory and
evidence, North Carolina State University and University of Colorado Working Paper.
Berndt, Antje, and Anastasiya Ostrovnaya, 2008, Do equity markets favor credit market news over options market
news?, North Carolina State University and Carnegie Mellon University Working Paper.
Bharath, Sreedhar T, and Tyler Shumway, 2008, Forecasting default with the Merton distance to default model,
Review of Financial Studies 21, 13391369.
Borensztein, Eduardo, Kevin Cowan, and Patricio Valenzuela, 2013, Sovereign ceilings lite? The impact of sovereign
ratings on corporate ratings, Journal of Banking and Finance 37, 40144024.
Bushman, Robert M, Joseph D Piotroski, and Abbie J Smith, 2004, What determines corporate transparency?,
Journal of Accounting Research 42, 207252.
Campbell, John Y, Jens Hilscher, and Jan Szilagyi, 2008, In search of distress risk, The Journal of Finance 63,
28992939.
Carr, Peter, and Liuren Wu, 2009, Variance Risk Premiums, Review of Financial Studies 22, 13111341.
Chava, Sudheer, Rohan Ganduri, and Chayawat Ornthanalai, 2012, Are credit ratings still relevant?, Georgia Institute
of Technology and University of Toronto Working Paper.
Choi, Stephen J, Mitu Gulati, and Eric A Posner, 2011, Pricing terms in sovereign debt contracts: a Greek case
study with implications for the European crisis resolution mechanism, Capital Markets Law Journal 6, 163187.
Claessens, Stijn, Daniela Klingebiel, and Sergio L Schmukler, 2007, Government bonds in domestic and foreign
currency: The role of institutional and macroeconomic factors, Review of International Economics 15, 370413.
Collin-Dufresne, Pierre, Robert S Goldstein, and J Spencer Martin, 2001, The determinants of credit spread changes,
The Journal of Finance 56, 21772207.
Coval, Joshua D, and Tobias J Moskowitz, 2001, The geography of investment: Informed trading and asset prices,
Journal of Political Economy 109, 811841.
Desai, Mihir A, C Fritz Foley, and James R Hines, 2004, A multinational perspective on capital structure choice and
internal capital markets, The Journal of Finance 59, 24512487.
Dieckmann, Stephan, and Thomas Plank, 2012, Default risk of advanced economies: An empirical analysis of credit
default swaps during the financial crisis, Review of Finance 16, 903934.
Dittmar, Robert F., and Kathy Yuan, 2008, Do sovereign bonds benefit corporate bonds in emerging markets?,
Review of Financial Studies 21, 19832014.
Djankov, Simeon, Caralee McLiesh, and Andrei Shleifer, 2007, Private credit in 129 countries, Journal of Financial
Economics 84, 299329.
Ducruet, Cesar, and Theo Notteboom, 2012, The worldwide maritime network of container shipping: Spatial structure
and regional dynamics, Global Networks 12, 395423.
Dunning, John H, 2002, Global Capitalism, FDI and competitiveness . vol. 2 (Edward Elgar Publishing).
Durbin, Erik, and David Ng, 2005, The sovereign ceiling and emerging market corporate bond spreads, Journal of
International Money and Finance 24, 631649.
Ejsing, Jacob, and Wolfgang Lemke, 2011, The Janus-headed salvation: Sovereign and bank credit risk premia during
2008-09, Economics Letters 110, 2831.
Elkamhi, Redouane, Kris Jacobs, and Xuhui Pan, 2012, The cross-section of recovery rates and default probabilities
implied by credit default swap spreads, Journal of Financial and Quantitative Analysis (forthcoming).
Ericsson, Jan, Kris Jacobs, and Rodolfo Oviedo, 2009, The determinants of credit default swap premia, Journal of
Financial and Quantitative Analysis 44, 109132.
Flannery, Mark J, Joel F Houston, and Frank Partnoy, 2010, Credit default swap spreads as viable substitutes for
credit ratings, University of Pennsylvania Law Review pp. 20852123.
Giannetti, Mariassunta, and Yishay Yafeh, 2012, Do cultural differences between contracting parties matter? Evi-
dence from syndicated bank loans, Management Science 58, 365383.
Grinblatt, Mark, and Matti Keloharju, 2001, How distance, language, and culture influence stockholdings and trades,
Guiso, Luigi, Paola Sapienza, and Luigi Zingales, 2009, Cultural biases in economic exchange?, The Quarterly Journal
of Economics 124, 10951131.
Hull, John, Mirela Predescu, and Alan White, 2004, The relationship between credit default swap spreads, bond
yields, and credit rating announcements, Journal of Banking and Finance 28, 27892811.
Jarrow, Robert, 2001, Default parameter estimation using market prices, Financial Analysts Journal 57, 7592.
La Porta, Rafael, Florencio Lopez de Silanes, Andrei Shleifer, and Robert Vishny, 2002, Investor protection and
corporate valuation, The Journal of Finance pp. 11471170.
La Porta, Rafael, Florencio Lopez de Silanes, Andrei Shleifer, and Robert W Vishny, 1997, Legal determinants of
external finance, The Journal of Finance 52, 11311150.
, 1998, Law and finance, Journal of Political Economy 106, 1113.

Lee, Jongsub, Andy Naranjo, and Stace Sirmans, 2014, CDS momentum: Slow-moving credit ratings and cross-
market spillovers, University of Florida Working Paper.
Lins, Karl V, Deon Strickland, and Marc Zenner, 2005, Do non-US firms issue equity on US stock exchanges to relax
capital constraints?, Journal of Financial and Quantitative Analysis 40, 109133.
Longstaff, Francis A, Sanjay Mithal, and Eric Neis, 2005, Corporate yield spreads: Default risk or liquidity? New
evidence from the credit default swap market, The Journal of Finance 60, 22132253.
Longstaff, Francis A, Jun Pan, Lasse H Pedersen, and Kenneth J Singleton, 2011, How sovereign is sovereign credit
risk?, American Economic Journal: Macroeconomics 3, 75103.
Merton, Robert C, 1974, On the pricing of corporate debt: The risk structure of interest rates, The Journal of
Finance 29, 449470.
Mian, Atif, 2006, Distance constraints: The limits of foreign lending in poor economies, The Journal of Finance 61,
14651505.
Miller, Darius P, and Natalia Reisel, 2012, Do country-level investor protections affect security-level contract design?
Evidence from foreign bond covenants, Review of Financial Studies 25, 408438.
Nini, Greg, David Smith, and Amir Sufi, 2012, Creditor control rights, corporate governance, and firm value, Review
of Financial Studies 25, 17131761.
Noe, Thomas H, 1998, Creditor rights and multinational capital structure, University of Oxford Working Paper.
Packer, Frank, and Haibin Zhu, 2005, Contractual terms and CDS pricing, BIS Quarterly Review pp. 89100.
Pagano, Marco, Ailsa A Roell, and Josef Zechner, 2002, The geography of equity listing: why do companies list
abroad?, The Journal of Finance 57, 26512694.
Pan, Jun, and Kenneth Singleton, 2008, Default and recovery implicit in the term structure of sovereign CDS spreads,
Peter, Marcel, and Mart

an Grandes, 2005, How important is sovereign risk in determining corporate default premia?
The case of south africa, IMF Working Paper No. 05/217.
Qian, Jun, and Philip E Strahan, 2007, How laws and institutions shape financial contracts: The case of bank loans,
Qiu, Jiaping, and Fan Yu, 2012, Endogenous liquidity of credit derivatives, Journal of Financial Economics 103,
611631.
Shleifer, Andrei, and Daniel Wolfenzon, 2002, Investor protection and equity markets, Journal of Financial Economics
66, 327.
Siegel, Jordan I, Amir N Licht, and Shalom H Schwartz, 2011, Egalitarianism and international investment, Journal
of Financial Economics pp. 122.
Tang, Dragon Yongjun, and Hong Yan, 2007, Liquidity and credit default swap spreads, University of Hong Kong
and University of South Carolina Working Paper.
Woolley-Meza, Olivia, Christian Thiemann, Daniel Grady, Jake Jungbin Lee, Hanno Seebens, Bernd Blasius, and
Dirk Brockmann, 2011, Complexity in human transportation networks: a comparative analysis of worldwide air
transportation and global cargo-ship movements, The European Physical Journal B 84, 589600.
Table 1. Summary Statistics

This table presents sovereign- and firm-level summary statistics for the 54 countries and 2,364 firms in our sample. 5-year Credit
Default Swap (CDS) spreads, reported in basis points (bps), are provided by Markit. N represents country-year observations at
the sovereign-level and firm-year observations at the firm-level. For a more detailed description of each variable, see our Internet
Appendix A.3.
Panel A. Sovereign-Level Summary Statistics

Units Mean Stan. Dev. Median Min Max N
Country Fundamentals:
Sovereign CDS Spread bps 122.330 246.630 56.915 1.000 3148.804 361
Ln(Sovereign CDS Spread) Ln(bps) 3.727 1.674 4.048 0.000 8.055 361
Sovereign CDS Depth number 7.428 3.376 7.351 2.000 16.103 361
Sovereign CDS Recovery Rate % 32.856 6.969 34.006 18.265 72.684 361
Sovereign S&P Credit Rating number 17.173 3.959 18.000 1.000 21.000 329
Region Sovereign CDS Spread bps 162.411 183.185 133.696 1.326 1531.556 361
Ln(Region Sovereign CDS Spread) Ln(bps) 4.096 1.581 4.296 0.129 7.499 361
Stock Market Volatility decimal 0.289 0.128 0.267 0.078 0.853 344
Ln(GDP) Ln($,billions) 6.013 1.352 5.799 2.513 9.603 357
Ln(GDP per Capita) Ln($) 9.796 1.091 10.175 6.477 11.627 357
Government Debt-to-GDP % 55.101 34.160 49.015 4.408 222.535 355
External Debt-to-GDP % 0.713 1.282 0.385 -5.486 10.378 344
Institutional & Informational Variables:
Property Rights 0.000 1.000 0.145 -2.099 1.267 360
Rule of Law 0.000 1.000 0.169 -2.430 1.141 341
Repudiation Risk 0.000 1.000 0.469 -1.867 1.354 341
Expropriation Risk 0.000 1.000 0.543 -2.571 1.062 289
Creditor Rights 0.000 1.000 -0.050 -1.932 1.833 342
Ln(Contract Enforcement Days) 0.000 1.000 0.256 -2.069 2.299 336
Disclosure Requirements: No. of Items Reported 0.000 1.000 0.270 -1.993 1.653 274
Disclosure Requirements: Reporting Frequency 0.000 1.000 0.231 -2.718 1.249 274
Panel B. Firm-Level Summary Statistics

Units Mean Stan. Dev. Median Min Max N
Firm Fundamentals:
Firm CDS Spread bps 195.511 468.413 85.828 4.065 8972.103 13204
Ln(Firm CDS Spread) Ln(bps) 4.483 1.180 4.452 1.403 9.102 13204
Firm CDS Depth number 5.628 4.125 4.131 2.000 27.367 13204
Firm CDS Recovery Rate % 37.042 5.204 39.481 2.830 95.720 13204
Ln(Firm CDS Recovery Rate) Ln(%) 3.600 0.161 3.676 1.043 4.561 13204
Firm S&P Credit Rating number 8.564 3.169 8.000 1.000 21.000 9370
Size Ln($) 23.634 2.360 23.366 10.108 32.626 13204
Leverage fraction 0.299 0.169 0.289 0.000 0.727 13204
Short-term Debt / Total Debt fraction 0.495 0.252 0.493 0.000 1.000 13204
Cash Flow / Total Assets fraction 0.077 0.056 0.070 -0.017 0.202 13204
Excess Stock Return decimal 0.032 0.394 -0.009 -1.913 2.204 11850
Stock Return Volatility decimal 0.048 0.026 0.042 0.000 0.174 11850
EDFMerton decimal 0.199 0.338 0.001 0.000 0.935 10716
Profit Margin fraction 0.208 0.177 0.169 -0.219 0.756 12104
Asset Tangibility fraction 0.293 0.248 0.240 0.000 0.977 12340
Foreign Assets / Total Assets fraction 0.183 0.243 0.067 0.000 1.000 13204
Foreign Sales / Total Sales fraction 0.292 0.296 0.212 0.000 1.000 13167
Number of Stock Exchanges number 2.348 1.087 2.000 1.000 8.000 13148
Number of Geographic Segments number 3.693 2.337 3.000 1.000 10.000 11966
Scaled Net Exposure & Extra Disclosure:
Ln(GDP per Capita) 0.000 1.000 -0.494 -1.085 8.259 12223
Stock Market Volatility 0.000 1.000 0.122 -11.267 12.978 12363
Property Rights 0.000 1.000 -0.320 -6.368 7.186 12240
Rule of Law 0.000 1.000 -0.339 -6.267 6.449 11663
Repudiation Risk 0.000 1.000 -0.439 -5.164 5.910 11569
Expropriation Risk 0.000 1.000 0.033 -7.935 8.691 11653
Creditor Rights 0.000 1.000 -0.031 -4.976 6.576 11618
Ln(Contract Enforcement Days) 0.000 1.000 0.245 -13.625 9.104 11141
Disclosure Req.: No. of Items Reported 0.000 1.000 -0.321 -0.321 8.743 12861
Disclosure Req.: Reporting Frequency 0.000 1.000 -0.282 -0.282 10.913 12867
Exodus from Sovereign Risk
Table 2. Determinants of Sovereign CDS Spreads
This table presents the determinants of sovereign Credit Default Swap (CDS) spreads for the 54 countries in our sample during 20042011. The dependent variable is the natural logarithm
of the annual mean sovereign CDS spread in basis points (bps). Regressions are at the country level using OLS with standard errors adjusted for clustering by country. Fixed effects (not
shown) include year, CDS restructuring type, and CDS currency. The variables in the categories of Property Rights Protection, Creditor Rights Protection, and Disclosure Requirements
are standardized such that the mean is equal to zero and the standard deviation is one. T-statistics are given in parentheses; statistical significance is indicated by *, **, and *** for 10%,
5%, and 1%, respectively.
Dependent Variable: Ln(Sovereign CDS Spread, bps)
(1) (2) (3) (4) (5) (6) (7) (8)

Country Fundamentals & Liquidity Effects
Ln(Region Sovereign CDS Spread) 0.372*** 0.365*** 0.349*** 0.344*** 0.427*** 0.375*** 0.378*** 0.375***
(9.35) (8.76) (8.63) (8.49) (9.41) (8.28) (8.08) (7.87)
Stock Market Volatility 1.336*** 1.904*** 1.591*** 2.942*** 1.877*** 1.878*** 2.982*** 3.069***
(3.66) (4.19) (3.59) (5.86) (4.54) (4.79) (5.29) (5.30)
Ln(GDP) -0.367*** -0.261*** -0.229*** -0.0977* -0.464*** -0.443*** -0.375*** -0.369***
(-7.33) (-5.59) (-5.44) (-1.87) (-6.14) (-5.58) (-5.07) (-4.81)
Government Debt-to-GDP 0.0117*** 0.00988*** 0.0109*** 0.00901*** 0.0111*** 0.0131*** 0.0103*** 0.0115***
(6.32) (5.53) (6.54) (5.69) (4.90) (5.77) (4.50) (5.08)
External Debt-to-GDP 0.0942*** 0.0825*** 0.0980*** 0.137*** 0.0530* 0.0676** 0.0675** 0.0497
(3.20) (2.76) (3.31) (4.57) (1.67) (2.24) (2.19) (1.61)
Sovereign CDS Depth 0.0826*** 0.0767*** 0.0875*** 0.0658*** 0.0992*** 0.101*** 0.115*** 0.134***
(5.21) (4.77) (5.87) (4.12) (5.49) (5.48) (5.77) (6.93)
Institutional Variables: Property Rights
Property Rights -0.654***
(-9.21)
Rule of Law -0.639***
(-8.77)
Repudiation Risk -0.701***
(-10.47)
Expropriation Risk -0.763***
(-11.03)
Institutional Variables: Creditor Rights
Creditor Rights -0.154*
(-1.78)
Ln(Contract Enforcement Days) 0.370***
(3.98)
Informational Variables:
Disclosure Requirements: No. of Items Reported -0.251***
(-2.96)
Disclosure Requirements: Reporting Frequency -0.0374
(-0.46)
Fixed Effects Year, CDS Restructuring Type, CDS Currency
Estimation Technique OLS with Standard Errors Clustered by Country
Observations 350 321 321 267 326 320 252 252
Adjusted R2 0.863 0.853 0.862 0.895 0.774 0.801 0.854 0.843
T-statistics are given in parentheses; Statistical significance indicated by *, **, and *** for 10%, 5%, and 1%, respectively.
Page 36
Table 3. CDS Spreads and Multinational Firms

This table presents average firm CDS spreads (in basis points) for our 2,364 firms in 54 countries at varying degrees of international
exposure (for a total of 13,204 firm years).
Panel A provides the results for various measures of aggregate international exposure. Each variable is divided into four buckets.
The first bucket, labeled Zero, represents firms that have zero international exposure (i.e., firms that are purely domestic). The
second, third and fourth buckets, labeled Low, Med and High, respectively, represent firms with increasing international
exposure. The buckets for Number of Stock Exchanges include Zero (1), Low (2-3), Medium (4-5), and High (6+). The buckets for
Number of Geographic Segments include Zero (1), Low (2-3), Medium (4-6), and High (7+). The buckets for Foreign Assets/Total
Assets include Zero (0%), Low (0.01%-15%), Medium (15%-40%), and High (40%-100%). The buckets for Foreign Sales/Total Sales
include Zero (0%), Low (0.01%-15%), Medium (15%-40%), and High (40%-100%). We also report a t-test for differences in means
between the High and Zero bucket groups.
Panel B reports firm exposures to foreign institutions by considering the Scaled Net Exposure variables, which are a weighted
average of the institutional values of the countries in which the firm has assets net of the institutional value of the firms home
country. Net Positive Exposure (Net Negative Exposure) is equal to one when a firms foreign assets are located in countries
with relatively stronger (weaker) institutions. We also report a t-test for differences in means between the positive and negative
exposure groups. The results are reported separately for the pre-crisis and crisis periods.
Panel A. Mean Corporate CDS Spreads and Aggregate International Exposure

Pre-Crisis Period (2004-2007) Crisis Period (2008-2011)
High Zero High Zero
Zero Low Med High Zero Low Med High
[t-Test] [t-Test]
No. of Stock Exchanges 106.051 78.143 45.420 27.158 -78.893*** 260.711 206.724 116.887 143.582 -117.129***
(140.65) (112.89) (73.46) (25.00) [4.52] (289.88) (239.63) (90.28) (121.32) [5.17]
N=1702 N=3614 N=357 N=302 N=2146 N=4235 N=419 N=373
No. of Geog. Segments 96.176 84.021 69.300 66.556 -29.620*** 275.135 223.911 167.676 124.000 -151.135***
(122.23) (114.14) (113.92) (67.99) [4.12] (311.89) (256.57) (173.35) (64.49) [4.81]
N=1242 N=1520 N=1749 N=899 N=1545 N=1743 N=1880 N=1388
Foreign Assets / Total Assets 87.500 81.400 68.613 64.951 -22.549*** 250.958 202.164 176.827 137.400 -113.558***
(132.73) (101.96) (113.35) (80.01) [4.91] (293.49) (224.32) (203.52) (96.7) [4.45]
N=2257 N=1373 N=1354 N=1032 N=2693 N=1700 N=1436 N=1359
Foreign Sales / Total Sales 88.739 78.241 68.944 74.348 -14.391*** 260.238 241.150 200.513 136.862 -123.376***
(125.37) (101.26) (134.9) (89.55) [3.81] (305.88) (259.02) (238.92) (97.45) [4.21]
N=1780 N=1479 N=1701 N=1038 N=2210 N=1546 N=1908 N=1505
Standard deviation in parentheses, t-statistic in brackets, statistical significance indicated by *, **, *** for 10%, 5%, 1%, respectively.
Panel B. Mean Corporate CDS Spreads and Exposure to Foreign Institutions

Pre-Crisis Period (2004-2007) Crisis Period (2008-2011)
Net Negative Net Positive Pos Neg Net Negative Net Positive Pos Neg
Exposure Exposure [t-Test] Exposure Exposure [t-Test]
Scaled Net Exposure: Property Rights
Property Rights 84.524 80.950 -3.574 238.588 215.052 -23.536***
(131.50) (119.11) [0.98] (284.88) (238.81) [3.48]
N=2060 N=1031 N=2722 N=963
Rule of Law 84.047 82.439 -1.608 235.790 222.854 -12.936**
(131.46) (121.67) [0.47] (280.25) (254.68) [1.97]
N=2204 N=693 N=2614 N=843
Repudiation Risk 91.636 69.819 -21.817*** 240.768 216.183 -24.585***
(138.7) (106.28) [4.66] (281.52) (253.25) [3.72]
N=2184 N=907 N=2792 N=893
Expropriation Risk 95.381 83.464 -11.917*** 236.900 220.031 -16.869**
(154.24) (127.94) [4.47] (276.96) (258.89) [2.53]
N=2431 N=633 N=2894 N=776
Scaled Net Exposure: Creditor Rights
Creditor Rights 91.011 65.736 -25.275*** 238.448 212.564 -25.884***
(136.01) (104.58) [3.57] (278.98) (248.84) [3.71]
N=1294 N=1596 N=1595 N=1890
Ln(Contract Enforcement Days) 60.001 90.675 30.674*** 191.908 241.832 49.924***
(99.84) (134.61) [4.49] (229.83) (280.19) [3.68]
N=1428 N=1633 N=1681 N=2003
Standard deviation in parentheses, t-statistic in brackets, statistical significance indicated by *, **, *** for 10%, 5%, 1%, respectively.
Table 4. Determinants of Corporate CDS Spreads
This table presents the determinants of corporate credit default swap (CDS) spreads for our sample of 2,364 companies in 54 countries during 2004-2011. The dependent
variable is the natural logarithm of the annual mean corporate CDS spread in basis points (bps). Key independent variables include the Scaled Net Exposure & Extra
Disclosure variables, which measure the firms asset exposure to foreign property rights and creditor rights as well as the firms extra disclosure required due to its equity
cross-listings. Panel A presents results of the baseline specification. Panel B shows robustness to various other types of international exposure and examines crisis period
interaction effects. CDS spreads are provided by Markit. Regressions are estimated at the firm-year level using OLS with standard errors adjusted for firm-level clustering.
For a more complete description of each variable, see Internet Appendix A.3. T-statistics are given in parentheses; statistical significance is indicated by *, **, and *** for
10%, 5%, and 1%, respectively.
Panel A. Corporate CDS Spreads and Exposure to Foreign Country Institutions

Baseline Specification Quantile Regression
(1) (2) (3) (4) (5)
Ln(Firm CDS Spread, bps) Firm CDS Spread, bps
Control Variables:
Ln(Sovereign CDS Spread, bps) 0.204*** 0.207*** 0.204*** 0.204***
(6.48) (6.50) (6.75) (6.19)
Sovereign CDS Spread, bps 0.461***
(5.50)
Size -0.0662*** -0.0650*** -0.0646*** -0.0646*** -3.621***
(-6.77) (-6.48) (-6.77) (-6.56) (-4.07)
Leverage 0.614*** 0.610*** 0.609*** 0.597*** 50.491***
(6.41) (6.21) (6.58) (6.14) (6.08)
Short-term Debt / Total Debt 0.0786 0.0903 0.0808 0.106 6.590
(1.20) (1.34) (1.26) (1.58) (1.32)
Cash Flow / Total Assets -0.168 -0.0487 -0.233 -0.0584 -14.980
(-0.90) (-0.26) (-1.29) (-0.31) (-0.89)
Excess Stock Return 0.0186 0.0129 0.0202 0.0182 1.753
(0.94) (0.62) (1.05) (0.88) (0.77)
Stock Return Volatility 2.362*** 2.369*** 2.298*** 2.373*** 336.462***
(23.06) (22.43) (22.13) (22.58) (14.11)
EDFMerton 0.455*** 0.423*** 0.451*** 0.424*** 42.522***
(11.62) (10.48) (11.45) (10.50) (7.22)
CDS Depth -0.0144*** -0.0141*** -0.0114*** -0.0112*** -0.629***
(-3.93) (-3.72) (-3.19) (-2.97) (-2.23)
Property Rights -0.0519*** -0.0392*** -5.884***
(-4.24) (-3.20) (-4.47)
Creditor Rights -0.0731*** -0.0554** -6.598***
(-2.71) (-2.02) (-3.18)
Disclosure Req.: No. of Items Reported -0.104*** -0.112*** -5.324***
(-6.77) (-7.39) (-3.78)
Fixed Effects Year, Industry, Country, CDS Restructuring Type, CDS Currency
Country Control Variables Ln(GDP per Capita), Government Debt-to-GDP, Stock Market Volatility
Standard Errors Cluster(Firm) Cluster(Firm) Cluster(Firm) Cluster(Firm) Block Bootstrap
Page 38
Observations 9091 8609 9695 8394 9415
Adjusted R2 0.611 0.610 0.609 0.616 0.260
T-statistics are given in parentheses; Statistical significance is indicated by *, **, and *** for 10%, 5%, and 1%, respectively.
Table 4 Continued

Panel B. Aggregate International Exposure, Foreign Fundamentals and Crisis Period Interactions
Dependent Variable: Ln(Firm CDS Spread, bps)
Aggregate International Exposure and Foreign Fundamentals Crisis Period Interactions

(1) (2) (3) (4) (5) (6) (7) (8) (9)
Aggregate International Exposure:
Foreign Assets / Total Assets -0.0268

(-0.42)
Foreign Sales / Total Sales 0.0719
(1.25)
Number of Stock Exchanges -0.123***
(-6.55)
Scaled Net Foreign Fundamentals:
Ln(GDP per Capita) -0.00927

(-0.59)
Stock Market Volatility 0.0557***
(4.04)
Property Rights -0.0368*** -0.0451*** -0.0293** -0.0357*** -0.0243** -0.0248 -0.0118

(-3.15) (-3.74) (-2.43) (-3.07) (-2.04) (-1.43) (-0.67)
Crisis Property Rights -0.0419** -0.0417**

(-2.20) (-2.14)
Creditor Rights -0.0551** -0.0568** -0.0511* -0.0533* -0.0389 -0.0350 -0.0190

(-2.01) (-2.08) (-1.91) (-1.92) (-1.41) (-1.13) (-0.61)
Crisis Creditor Rights -0.0474*** -0.0444***

(-3.01) (-2.74)
Disclosure Req.: No. of Items Reported -0.111*** -0.115*** -0.0483*** -0.110*** -0.109*** -0.0941*** -0.0996***
(-7.23) (-7.51) (-2.78) (-7.03) (-7.07) (-5.85) (-6.22)
Crisis Disclosure Req.: No. of Items Reported -0.0243** -0.0310**

(-2.16) (-2.37)
Firm Control Variables Size, Leverage, Short-term Debt/Total Debt, Cash Flow/Total Assets, Excess Stock Return,
Stock Return Volatility, EDF Merton, CDS Depth
Country Control Variables Ln(Sovereign CDS Spread, bps), Ln(GDP per Capita), Government Debt-to-GDP, Stock Market Volatility
Estimation Technique OLS with Standard Errors Clustered by Firm
Observations 8385 8394 8276 8269 8385 9091 8609 9695 8394
Adjusted R2 0.616 0.624 0.615 0.616 0.616 0.611 0.610 0.609 0.617
Page 39
Table 5. Corporate Credit Ratings and Exposure to Foreign Institutions

This table presents the determinants of corporate credit ratings for 1,374 companies in 43 countries over the years 2004-2011. The
dependent variable is a firms S&P credit rating converted to a numerical score in which a higher number corresponds to a higher
credit quality (for e.g., 21=AAA, 20=AA+, etc.). Key independent variables include the Scaled Net Exposure variables, which
measure the firms asset exposure to foreign property rights and creditor rights as well as the firms extra disclosure required due to
its equity cross-listings. Column (7) includes additional controls not shown: Ln(Sovereign CDS Spread, bps), CDS Depth, and CDS
Restructuring Type/Base Currency dummies. T-statistics are given in parentheses; statistical significance is indicated by *, **, and
*** for 10%, 5%, and 1%, respectively.
Dependent Variable: Firm S&P Credit Rating Ln(Firm CDS Spread, bps)
(1) (2) (3) (4) (5) (6) (7)
Firm Fundamentals:
Sovereign S&P Credit Rating 0.443*** 0.430*** 0.528*** 0.426*** 0.439*** 0.484***
(6.90) (6.52) (7.44) (5.85) (6.50) (5.87)
Size 0.275*** 0.269*** 0.296*** 0.255*** 0.268*** 0.257***

(6.58) (6.53) (6.08) (6.34) (6.82) (5.67)
Leverage -2.380*** -2.292*** -2.448*** -2.437*** -2.474*** -2.478***

(-5.62) (-5.44) (-5.48) (-5.85) (-6.00) (-5.43)
Short-term Debt / Total Debt -0.261 -0.219 -0.531 -0.373 -0.379 -0.689**
(-0.93) (-0.78) (-1.63) (-1.30) (-1.38) (-2.15)
Cash Flow / Total Assets 1.655** 1.444* 1.156 1.422* 1.987*** 1.222
(2.16) (1.89) (1.33) (1.84) (2.66) (1.47)
Excess Stock Return 0.0403 0.0417 0.0282 0.0121 0.0341 0.0277

(0.55) (0.57) (0.32) (0.16) (0.47) (0.31)
Stock Return Volatility -8.492*** -8.536*** -8.102*** -8.345*** -8.371*** -8.001***

(-21.36) (-21.66) (-17.77) (-20.62) (-21.42) (-18.00)
EDFMerton -0.252 -0.235 -0.264 -0.317** -0.280* -0.396**

(-1.62) (-1.51) (-1.55) (-2.03) (-1.83) (-2.35)
Firm S&P Credit Rating -0.204***

(-39.85)
Aggregate Foreign Exposure:
Foreign Assets / Total Assets 0.974***

(3.95)
Foreign Sales / Total Assets 1.415***

(5.46)
Property Rights -0.136 0.124 -0.0351***

(-0.83) (0.88) (-3.59)
Creditor Rights 0.353 0.404 -0.0397**

(1.32) (1.47) (-2.12)
Disclosure Req.: No. of Items Reported 0.487*** 0.553*** -0.00933

(3.05) (3.17) (-0.87)
Fixed Effects Year, Industry, Country
Country Controls Ln(GDP per Capita), Government Debt-to-GDP, Stock Market Volatility
Observations 7988 7970 7459 7040 7932 6902 7789
Adjusted R2 0.442 0.448 0.431 0.454 0.456 0.468 0.764
Table 6. Determinants of Corporate CDS Recovery Rates

This table presents the determinants of corporate CDS recovery rates for our sample of 2,364 companies in 54 countries during 2004-
2011. The dependent variable is the natural logarithm of the annual mean corporate CDS recovery rate in percentage points (%).
Key independent variables include the Scaled Net Exposure variables, which measure the firms foreign asset exposure to property
rights and creditor rights as well as the firms extra disclosure required through its equity cross-listings. CDS Recovery rate data
are acquired from Markit. Industry Asset Specificity, Industry Q, and Industry Distress Dummy variables are constructed following
Acharya, Bharath and Srinivasan (2007). Industry Asset Specificity is the industry median ratio of machinery and equipment to
total assets. Industry Q is the industry median ratio of a firms market value to book value. Industry Distress Dummy is flagged as
one if the industry median stock return is less than 30%. For a more detailed description of each variable, see Internet Appendix
A.3. Regressions are estimated using OLS with standard errors adjusted for clustering by firm. T-statistics are given in parentheses;
statistical significance is indicated by *, **, and *** for 10%, 5%, and 1%, respectively.
Dependent Variable: Ln(Firm CDS Recovery Rate, %)
(1) (2) (3) (4)

Firm, Industry and Country Control Variables:
Ln(Sovereign CDS Recovery Rate) 0.118*** 0.139*** 0.134*** 0.101**

(3.10) (4.73) (4.77) (2.49)
Size 0.00253* 0.00261* 0.00319** 0.00199

(1.69) (1.86) (2.38) (1.26)
Leverage -0.00752 -0.0198 -0.0221 -0.00890

(-0.50) (-1.25) (-1.47) (-0.56)
Profit Margin 0.0231 0.0411** 0.0371** 0.0242

(1.47) (2.54) (2.38) (1.47)
Asset Tangibility 0.0433*** 0.0345*** 0.0329*** 0.0444***

(4.04) (3.31) (3.29) (3.98)
Industry Asset Specificity -0.0448*** -0.0515*** -0.0486*** -0.0434**

(-2.78) (-3.14) (-3.23) (-2.40)
Industry Q 0.00899*** 0.0109*** 0.00967*** 0.0114***

(4.09) (5.07) (4.95) (4.90)
Industry Distress Dummy -0.0151*** -0.0217*** -0.0157*** -0.0220***

(-2.58) (-4.10) (-3.16) (-3.32)
Property Rights 0.0161* 0.0199**

(1.74) (2.10)
Creditor Rights 0.00793** 0.0117***

(2.21) (2.77)
Disclosure Req.: No. of Items Reported 0.00194 0.00254

(1.01) (1.21)
Country Control Variables Ln(GDP per Capita), Government Debt-to-GDP, Stock Market Volatility
Observations 9826 9228 10364 9096
Adjusted R2 0.407 0.411 0.398 0.413
Table 7. Further Robustness Checks: Selection and Reverse Causality
This table presents further robustness checks for potential selection and reverse causality concerns in our baseline results in Table 4. T-statistic is in parentheses; statistical
significance is indicated by *, **, and *** for 10%, 5%, and 1%, respectively.
Panel A presents the results of (i) adding firm fixed effects to the baseline specification (Models 14), (ii) separating out the differential effect between net positive and
net negative institutional exposures (Models 5 and 6), and (iii) testing the added effect of delisting when it reduces the firms overall required level of disclosure (Model 7).
The variable Net Positive (Net Negative) indicates a firms Scaled Net Exposure that is constructed exclusively using its foreign asset exposure in stronger (weaker)
institutions than the firms home country (i.e., ScaledN etExposurei,t = N etP ositivei,t N etN egativei,t ). We construct these variables for foreign property rights (Model
5) and creditor rights (Model 6). These variables are standardized to have a mean of zero and standard deviation of one. In Model 7, the dummy variable Delisted from any
exchange indicates that the firm delisted from an equity exchange during the year. Delisted from high disclosure exchange dummy indicates that the firm delisted from a
relatively strict disclosure exchange that reduced the firms maximum disclosure requirement. Other fixed effects, firm control variables, and country control variables are as
listed in the table.
Panel B presents the 2SLS estimation results of using a dummy variable indicating whether a firm has exposure to foreign countries that experience large negative shocks
in property rights as an instrument for the firms Scaled Net Exposure to foreign property rights. The dependent variable in Stage 1 is the firms Scaled Net Exposure:
Property Rights, and the dependent variable in Stage 2 is the natural logarithm of the firms mean annual CDS spread in bps. We report results from five models in
Panel B. The model ALL NEGATIVE includes exposure to all negative changes in property rights (with a drop of at least 10 out of 100) as the instrument for Scaled Net
Exposure. For this model, we further report results using one- and two-year lagged asset allocations prior to each negative property rights shock. The model DISTANCE
includes only the subset of ALL NEGATIVE shocks in countries that are farther than 4,000 miles away from the firm. The model CULTURE includes only the subset of
ALL NEGATIVE shocks in countries that are culturally distinct from the firms home country. Culturally distinct is defined as being more than one standard deviation
away in both the Traditional/Secular value and the Survival/Self-expression value of the World Values Survey. Starting with the subset of ALL NEGATIVE shocks, the
model LANGUAGE excludes shocks that occur in countries with the same primary language as the firms home country and also excludes all shocks in countries with English,
French, German, or Spanish primary languages. The model COUNTRY EVENTS includes only the subset of ALL NEGATIVE shocks for a select group of countries that
experience unanticipated nationalizations or eliminations of democratic processes. In consultation with the Heritage Foundation, the selected countries are Argentina in
2009, Egypt in 2007, Guinea in 2009, and Venezuela in 2008. For in-depth information on events occurring in these countries, see the Internet Appendix A.8. Fixed effects,
firm control variables, and country control variables are as listed in the table. For more information on the property rights index produced by the Heritage Foundation, see
http://www.heritage.org/index/property-rights.
Page 42
Table 7 Continued

Panel A. Firm Fixed Effects, Asymmetric Institutional Exposure and Stock Delisting Effects
Dependent Variable: Ln(Firm CDS Spread, bps)
Firm Fixed Effects Asymmetric Exposure Delisting

(1) (2) (3) (4) (5) (6) (7)
Property Rights -0.0290** -0.0244**
(-2.57) (-2.16)
Creditor Rights -0.0569** -0.0427*
(-2.32) (-1.68)
Disclosure Req.: No. of Items Reported -0.0398*** -0.0464***
(-3.11) (-3.48)
Exposure to Foreign Property Rights:
Net Positive (Stronger than home country) -0.0687**
(-2.13)
Net Negative (Weaker than home country) 0.0558*
(1.90)
Exposure to Foreign Creditor Rights:
Net Positive (Stronger than home country) -0.0981**
(-1.99)
Net Negative (Weaker than home country) 0.0364*
(1.72)
Delisting Effects:
Delisted from any exchange 0.0419*
(1.81)
Delisted from high disclosure exchange 0.217**
(2.03)
Observations 9091 8609 9695 8394 9091 8609 8806
Adjusted R2 0.683 0.682 0.679 0.683 0.613 0.612 0.625
Page 43
Table 7 Continued

Model:
ALL NEGATIVE All large negative property rights shocks ( 10 out of 100).
DISTANCE Subset of ALL NEGATIVE that are geographically distant (Threshold: 4000 miles away).
CULTURE Subset of ALL NEGATIVE that are culturally different (Threshold: 1 standard deviation of World Values Survey cultural measure).
LANGUAGE Subset of ALL NEGATIVE that are in a different language (while also excluding English, French, German, and Spanish shocks altogether).
COUNTRY EVENTS Subset of ALL NEGATIVE for a select group of countries that experience unanticipated nationalizations or eliminations of democratic processes.
(In consultation with the Heritage Foundation, the countries are Argentina in 2009, Egypt in 2007, Guinea in 2009, and Venezuela in 2008.)
Panel B. Addressing Reverse Causality: Exposure to Large Property Rights Shocks as Instrument for Scaled Net Exposure
Lagged Asset Allocation

Baseline One Year Prior Two Years Prior Additional Robustness Checks
ALL ALL ALL COUNTRY
NEGATIVE NEGATIVE NEGATIVE DISTANCE CULTURE LANGUAGE EVENTS
Stage 1: Scaled Net Exposure: Property Rights (1) (2) (3) (4) (5) (6) (7)
Foreign Property Rights Shock:
Negative Exposure Dummy -0.131*** -0.178*** -0.161*** -0.184*** -0.134*** -0.174*** -0.0763**
(-7.61) (-6.09) (-5.46) (-4.29) (-5.23) (-8.43) (-1.99)
Adjusted R2 0.195 0.177 0.146 0.190 0.188 0.192 0.187
F-Statistic 57.9 37.1 29.8 18.4 27.35 71.1 4.0
[P-Value] [0.0000] [0.0000] [0.0000] [0.0000] [0.0000] [0.0000] [0.0470]
Stage 2: Ln(Firm CDS Spread, bps) (1) (2) (3) (4) (5) (6) (7)
Scaled Net Exposure:
Property Rights (Instrumented) -1.006** -1.010* -0.833* -0.987* -1.435*** -0.817** -1.052**
(-2.22) (-1.85) (-1.82) (-1.86) (-2.87) (-2.32) (-2.18)
Adjusted R2 0.610 0.585 0.599 0.604 0.603 0.605 0.601
Estimation Technique Two Stage Least Squares
Firm-Year Shocks 426 323 277 223 140 270 63
Observations 9091 7901 6194 9091 9091 9091 9091
Page 44
Figure 1. Global Sovereign CDS Spreads

This figure presents the time series of the GDP-weighted and equally-weighted mean of global 5-year sovereign CDS daily spreads
(in basis points) of 54 countries over the years 2004-2011. CDS data are provided by Markit. Numbers are placed on the graph to
mark important events and outline the initial rising credit risk of the financial sector, government bailouts that followed, and the
concurrent increase in sovereign credit risk across the globe. 33
1. July 31, 2007. Bear Stearns liquidates hedge funds.
2. September 15, 2008. Lehman Bros. files for Ch. 11 bankruptcy.
3. October 3, 2008. US congress passes Troubled Asset Relief Fund (TARP).
4. December 16, 2009. S&P cuts Greeces credit rating.
5. April 27-28, 2010. S&P cuts credit rating of Greece, Portugal and Spain.
6. May 2-9, 2010. Greece accepts bailout; European Financial Stability Facility (EFSF) created.
7. November 28, 2010. Spain accepts bailout.
8. May 17, 2011. Portugal accepts bailout.
9. September 18 October 13, 2011. S&P cuts credit rating of Spain, Italy, and 24 Italian banks.
33
For a more complete timeline of the recent financial crisis, visit http://timeline.stlouisfed.org.
Figure 2. Global Sovereign Ceiling Violations (SCVs) in the CDS Market

As some visual motivation to provide evidence on firms delinking from their sovereign risks, the top figure in Panel A presents the
daily frequency of global sovereign ceiling violations (SCVs) in the credit default swap (CDS) market as measured by the percentage
of 5-year CDS contracts that are classified as SCVs using the simple method (i.e., the firm CDS daily spread is less than that of
its sovereign counterpart with equal contractual terms). CDS spread data are retrieved from Markit and include 2,364 firms in 54
countries over 2004-2011. The bottom figure of Panel A presents the daily magnitude of SCVs in the global CDS market during the
2008-2011 crisis period using the mean and median of the firm CDS spread minus sovereign CDS spread, which are computed each
day using observations that are classified as SCVs using the simple method. In Panel B, the top and bottom figures compare the
quantity of sovereign ceiling violations (SCVs) in the 5-year credit default swap (CDS) market in countries with strong and weak
institutions during the pre-crisis period (top) and the crisis period (bottom), respectively. We have three categories of institutional
variables: property rights protection (Property Rights, Rule of Law, Repudiation Risk, Expropriation Risk ), creditor rights protection
(Creditor Rights, Ln(Contract Enforcement Days)), and disclosure requirements (Number of Items Reported, Reporting Frequency ).
See variable descriptions in Section 3.1 and Internet Appendix A.3 for more information on the institutional variables. For each
institutional variable, we divide countries into two groups: top 50% (strong) and bottom 50% (weak). We then compute the SCVs
in each group in each year as a percentage of all firm-year observations in the group. To determine whether a firm is a sovereign
ceiling violator in a given year, we use the mean method, which flags a firm as a violator in the given year if its annual average
5-year CDS daily spread is lower than that of its home country. See Internet Appendix A.5 for more details of the mean method
procedure. In this analysis in Panel B, we adjust for transaction costs by including bid-ask components for each country in each year.
The bid-ask information on the CDS data is retrieved from an alternative data source, Credit Market Aanlysis (CMA), provided
through Datastream.
Panel A. Daily Patterns of SCVs in the CDS Market
Global SCVs in the CDS Market, Transaction-cost unadjusted (Percentage of all contracts)
Magnitude of Global SCVs in the CDS Market during the Crisis Period (Mean and median of all SCVs)
Panel B. Cross-country Patterns of Annual SCVs in the CDS Market Before/After the Crisis
(Transaction-cost Adjusted)
Pre-Crisis (2004 2007)
Crisis (2008 2011)

Figure 3. Sovereign CDS Spreads and Country Institutional Characteristics
This figure presents scatterplots between the natural logarithm of the annualized average 5-year sovereign CDS daily spreads (in basis points) and various country institutional
characteristics. Panel A, Panel B, and Panel C examine three distinct categories of institutional variables: Property Rights Protection, Creditor Rights Protection, and
Disclosure Requirements. Each plot contains country-year observations. All institutional variables are standardized. Large values indicate high institutional strength
(Ln(Contract Enforcement Days) being the only exception). CDS spreads are provided by Markit and include 54 countries over 2004-2011.
Page 48
Figure 4. Asset Allocations to Countries with Large Negative Property Rights Shocks
This figure represents firms aggregate asset allocation (fraction of total assets) to countries that experience a large negative property
rights shock (at Year 0). A large negative shock is defined as a Heritage Foundation Property Rights Index downward change of at
least 10 (out of 100), which corresponds to less than 5% events in the left tail of the variables distribution. The aggregate asset
allocation represents the fraction of total assets allocated to these countries and is computed as follows:
(1) For each country, foreign firms asset allocation to the country is averaged equally across firms.
(2) The country-level asset allocations are then averaged equally across countries.
Firm-year level geographic segment data are provided by Worldscope.
.3
Country-Specific Allocation (Fraction of Total Assets)
.1 .15 .2 .25
-3 -2 -1 0 1
Event Year = 0

Exodus From Sovereign Risk - Global Asset and Informational Networks in The Pricing of Corporate Credit Risk

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Exodus From Sovereign Risk - Global Asset and Informational Networks in The Pricing of Corporate Credit Risk

Uploaded by

Copyright:

Available Formats

Exodus from Sovereign Risk:

Global Asset and Information Networks in the Pricing of Corporate

Jongsub Lee, Andy Naranjo, and Stace Sirmans

Department of Finance, Insurance & Real Estate

August 29, 2014

JEL Classification: F23, F34, F36, G01, G15, K40

Electronic copy available at: http://ssrn.com/abstract=2492204

Electronic copy available at: http://ssrn.com/abstract=2492204

1 Institutional Background on the CDS Market

3 Data and Variables

3.1 Sovereign-level Variables

To measure a countrys level of informational transparency, we utilize two disclosure requirement

Disclosure Requirements: Number of Items Reported records a countrys required number of

Disclosure Requirements: Reporting Frequency measures a countrys required timeliness of

3.2 Firm-level Variables

3.3 Scaled Net Exposure and Extra Disclosure Variables

market segment data.

ScaledN etExposurei,t =F orAsset%i,t

have a mean of zero and standard deviation of one.

4.1 Sovereign Risk: Local Institutional and Information Qualities

Ln(Sovereign CDS Spread )i,j,k,t = 0 + k + t

4.2 Exodus from Sovereign Risk: Global Asset Geographic Location

4.2.1 The Determinants of Corporate CDS Spreads

the two sub-periods and report the results in Panel A of Table 3.

contractual terms-l, and year-t:

Ln(Firm CDS Spread )i,j,k,l,t = 0 + t + j + k + l

In Panel B of Table 4, we provide some additional identification checks. Each specification in

Ln(Firm CDS Recovery Rate)i,j,k,l,t = 0 + t + j + k + l

4.2.3 Selection and Reverse Causality: Causal Evidence

, 1998, Law and finance, Journal of Political Economy 106, 1113.

Peter, Marcel, and Mart

Table 1. Summary Statistics

Panel A. Sovereign-Level Summary Statistics

Panel B. Firm-Level Summary Statistics

Dependent Variable: Ln(Sovereign CDS Spread, bps)

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

Table 3. CDS Spreads and Multinational Firms

Panel A. Mean Corporate CDS Spreads and Aggregate International Exposure

Panel B. Mean Corporate CDS Spreads and Exposure to Foreign Institutions

Panel A. Corporate CDS Spreads and Exposure to Foreign Country Institutions

Aggregate International Exposure and Foreign Fundamentals Crisis Period Interactions

Foreign Assets / Total Assets -0.0268

Ln(GDP per Capita) -0.00927

Property Rights -0.0368*** -0.0451*** -0.0293** -0.0357*** -0.0243** -0.0248 -0.0118

Crisis Property Rights -0.0419** -0.0417**

Creditor Rights -0.0551** -0.0568** -0.0511* -0.0533* -0.0389 -0.0350 -0.0190

Crisis Creditor Rights -0.0474*** -0.0444***

Crisis Disclosure Req.: No. of Items Reported -0.0243** -0.0310**

Table 5. Corporate Credit Ratings and Exposure to Foreign Institutions

Size 0.275*** 0.269*** 0.296*** 0.255*** 0.268*** 0.257***

Leverage -2.380*** -2.292*** -2.448*** -2.437*** -2.474*** -2.478***

Excess Stock Return 0.0403 0.0417 0.0282 0.0121 0.0341 0.0277

Stock Return Volatility -8.492*** -8.536*** -8.102*** -8.345*** -8.371*** -8.001***

EDFMerton -0.252 -0.235 -0.264 -0.317** -0.280* -0.396**

Firm S&P Credit Rating -0.204***

Foreign Assets / Total Assets 0.974***

Foreign Sales / Total Assets 1.415***

Property Rights -0.136 0.124 -0.0351***

Creditor Rights 0.353 0.404 -0.0397**

Disclosure Req.: No. of Items Reported 0.487*** 0.553*** -0.00933

Table 6. Determinants of Corporate CDS Recovery Rates

Property Rights -0.0368* -0.0451* -0.0293 -0.0357* -0.0243** -0.0248 -0.0118

Crisis Property Rights -0.0419 -0.0417

Creditor Rights -0.0551 -0.0568 -0.0511* -0.0533* -0.0389 -0.0350 -0.0190

Crisis Creditor Rights -0.0474* -0.0444*

Crisis Disclosure Req.: No. of Items Reported -0.0243 -0.0310

Size 0.275* 0.269* 0.296* 0.255* 0.268* 0.257*

Leverage -2.380* -2.292* -2.448* -2.437* -2.474* -2.478*

Stock Return Volatility -8.492* -8.536* -8.102* -8.345* -8.371* -8.001*

Disclosure Req.: No. of Items Reported 0.487* 0.553* -0.00933

Ln(Sovereign CDS Recovery Rate) 0.118* 0.139* 0.134* 0.101

Profit Margin 0.0231 0.0411 0.0371 0.0242

Asset Tangibility 0.0433* 0.0345* 0.0329* 0.0444*

Industry Asset Specificity -0.0448* -0.0515* -0.0486* -0.0434

Industry Q 0.00899* 0.0109* 0.00967* 0.0114*

Industry Distress Dummy -0.0151* -0.0217* -0.0157* -0.0220*

Creditor Rights 0.00793 0.0117*