Week 4

Finance Research Letters 15 (2015) 232–238

Contents lists available at ScienceDirect

Finance Research Letters

journal homepage: www.elsevier.com/locate/frl

Credit contagion and competitive effects of bond

rating downgrades along the supply chain

Jung-Hsien Chang a, Mao-Wei Hung b, Feng-Tse Tsai c,∗

a Department of Banking and Finance, National Chi Nan University, No. 1, University Road, Puli, Nantou County

54561, Taiwan b Department of International Business, National Taiwan University, No. 1, Sec. 4, Roosevelt Rd., Taipei City 106,

Taiwan c Department of Finance, Asia University, 500, Lioufeng Rd., Wufeng, Taichung 41354, Taiwan

a r t i c l e i n f o

Article history:

Received 26 August 2015

Accepted 16 October 2015

Available online 24 October 2015

JEL Classification:

G10

G14

G30

G33

Keywords:

Credit risk

Credit rating

Contagion effect

Competitive effect

Supply chain

a b s t r a c t

This study investigates credit risk effects of credit rating downgrades

on downgraded firms’ intra-industry rivals (horizontal relation), sup-

pliers and customers (vertical relation). Using event study approach,

we analyze credit default swap (CDS) spread changes for downgraded

firms’ rivals, suppliers and customers. The result shows that rivals

and suppliers experience significant credit spread increases during

bond rating downgrades. Suppliers suffer from vertical credit conta-

gion only conditioning on occurrence of horizontal credit contagion

(increasing credit risk among intra-industry rivals). When horizontal

credit competitiveness happens (decreasing credit risk among intra-

industry rivals), customers appear to have vertical credit competitive

effects.

© 2015 Elsevier Inc. All rights reserved.

1. Introduction

Firms are not independent entities in the economy but are linked to each other through explicit or

implicit relationships. Because of the economic dependency among the firms, any shock to one firm

has a rippling effect to its linked partners. The linkages between firms in the economy are of special

∗ Corresponding author. Tel.: +886 423323456; fax: +886 423321181. E-mail address: [email protected], [email protected] (F.-T. Tsai).

http://dx.doi.org/10.1016/j.frl.2015.10.006

1544-6123/© 2015 Elsevier Inc. All rights reserved.

J.-H. Chang et al. / Finance Research Letters 15 (2015) 232–238 233

i

i

(

f

s

e

h

u

s

g

c

t

a

p

(

b

e

fi

e

i

f

o

b

t

n

t

S

s

l

d

f

o

s

l

a

i

a

d

b

m

r

c

t

fi

i

S

nterest in the case of financial distress. One firm suffering from financial problems can have valuation

mplications for firms which are linked in the product market (industry rivals) or along the supply chain

suppliers and customers). Therefore, default clustering often happens because the economic links of de-

ault event firms spread credit risk to other related firms. Existing studies investigate the wealth effects of

tockholders for intra-industry rivals, suppliers and customers in financial distress events. However, little

vidence on the wealth effects of debtholders for distressed firms’ suppliers and customers. In addition,

ow these effects of suppliers and customers interact with the wealth effects for industry rivals is also

nknown in the literature. The goal of this paper is to close the gap in the related research.

This study is motivated by several strands of research in finance. One line of research examines

tock reaction of related firms during financial distress events such as bankruptcy of Chapter 11 (reor-

anization), bankruptcy of Chapter 7 (liquidation), pre-filing distress (dramatic stock price drop) and

redit rating downgrades (or negative outlook). A representative study by Lang and Stulz (1992) shows

hat bankruptcy announcements have negative effect (contagion effect) for highly levered industries

nd industries where nonbankrupt and bankrupt firms are highly correlated in stock returns, but have

ositive effect (competitive effect) for highly concentrated industries with low leverage. Jorion and Zhang

2010) find that stock returns of industry portfolios have contagion effects for investment-grade firms

ut have predominant competitive effects for speculative-grade firms. Additionally, Hertzel et al. (2008)

xtend previous research by examining the wealth effect of distress on suppliers and customers. They

nd that significant contagion effects of stockholders for suppliers exist in bankruptcy filings and these

ffects are more negative when intra-industry contagion is more severe.

Another line of research discusses the wealth effects of debtholders for intra-industry rivals dur-

ng financial distress events. Jorion and Zhang (2007) examine the intra-industry information trans-

er effects of credit events in the credit default swaps (CDS) and stock markets. They provide evidence

f credit contagion effects for Chapter 11 bankruptcies and credit competitive effects for Chapter 7

ankruptcies. Comparing to Lang and Stulz (1992), they find that reactions in stock markets are rela-

ively less significant than reactions in CDS markets. Jorion and Zhang (2009) find that bankruptcy an-

ouncements have credit contagion (negative abnormal stock returns and increasing CDS spreads) for

rade counterparties (creditors) and this counterparty default risk can result in clustering of default.

pillover effects of credit risk among countries in the Eurozone (e.g., Alemany et al., 2015) and between

tock/bond and CDS markets (e.g., Chang et al., 2012; Delis and Mylonidis, 2011) are also discussed in the

iterature.

To the best of my knowledge, no literature covers information transfer effect of credit events on

ebtholders along the supply chain. Current studies (e.g., Hertzel et al., 2008) only focus on the wealth ef-

ects of stockholders along the supply chain in credit events. However, it is intuitive that credit contagion

r competitive effects along the supply chain are more directly linked to these credit events. In addition,

everal credit events imply different reactions across stock and credit markets. For example, increasing

everage leads to a wealth transfer from debtholders to stockholders and hence results in distinct stock

nd credit reactions (Goh and Ederington, 1993). Therefore, the objective of this paper is to discover the

nformation transfer effects of credit events along the supply chain in credit markets.

Comparing with bankrupt events, credit rating downgrade events are more frequent, cover more firms

nd hence the effect of negative credit events along the supply chain can be more comprehensive. In ad-

ition, credit rating downgrades communicate credit quality deterioration of bond issuers to investors

efore the firms go bankruptcy.1 Thus, it is interesting to investigate if credit rating downgrade infor-

ation also transfers across market segments. We focus on the information transmission effect of credit

ating downgrade through the supply chain, i.e., suppliers and customers in product markets. The supply

hain relationship helps these connected firms inherit production efficiency. Therefore, when one firm in

he supply chain suffers from credit rating downgrade, the event may influence operation of its linked

rms and hence their stock returns or credit risk.

The remainder of this paper is organized as follows. Section 2 presents the data source and process-

ng as well as summary statistics. In Section 3, we illustrate the methodology applied in this research.

ection 4 summarizes our empirical results. We conclude in Section 5.

1 Fabozzi et al. (2007) take rating as one fundamental variable on the pricing of CDS.

234 J.-H. Chang et al. / Finance Research Letters 15 (2015) 232–238

Table 1

Sample characteristics.

Panel A: distribution of downgrade events by issuer (full sample, N = 358) Issuer # Mean Median Max Min SD

100 3.58 2.5 19 1 3.39

Panel B: industry distribution

1-digit SIC Industry Obs. Fraction (%)

1 Mining and construction 12 3.35

2 Manufacturing (food–petroleum) 72 20.11

3 Manufacturing (plastics–electronics) 170 47.49

4 Transportation 25 6.98

5 Wholesale trade and retail trade 47 13.13

6 Finance, insurance, and real estate 12 3.35

7 Services (hotel–recreation) 19 5.31

8 Services (health–private household) 1 0.28

Total 358 100

Panel C: distribution of rivals, suppliers, and customers

Mean Median Max Min SD

Rivals 8.46 5 44 1 8.11

Suppliers 6.65 4 27 1 7.99

Customers 1.65 1 6 1 1.07

2. Data

The research period covers from 2001 to 2008. The data come from several sources including Mergent

FISD, Markit, Compustat, and CRSP. Following Jorion et al. (2005), we limit the sample to taxable corporate

debts in United States and exclude Yankee debts and private debts. Issuer credit rating is based on its debt

(issue) credit rating (Odders-White and Ready, 2006). CDS spreads are compiled from Markit and we only

choose 5-year senior unsecured CDS spreads with Modified Restructure (MR) clauses.

Supply chain relationship is constructed using data from Compustat segment file. We follow Hertzel

et al. (2008) to identify suppliers and customers of rating event firms. This is based on Financial Account-

ing Standards (FAS131) in United States that public firms are required to provide information about their

(main) customers with sales over 10% of their total sales. To find the suppliers of rating event firms, we

identify all Compustat firms that list a rating downgrade firm as a major customer. Similarly, to form the

sample of rating downgrade firm customers, we reverse the above process and identify the firms listed by

rating downgrade firms as major customers.

The procedure of sample filtering is as follows. We only adopt Standard and Poor’s ratings and Moody’s

ratings and keep the largest rating change for the issuer at the same date since that rating adjustment has

largest impacts on the market. Only the first downgrade is adopted if Standard & Poor’s and Moody’s

assign downgrade ratings consecutively for the same issuer. Moreover, we require the issuer to have daily

CDS and stock returns 250 days before and 5 days after the downgrade announcement date in the Markit

and CRSP database respectively. Finally, we link downgraded issuers to their suppliers and customers

and keep only those samples having suppliers or customers. Therefore, the final sample constitutes 358

downgrade announcements, 100 downgraded issuers, 358 industry portfolios, 155 suppliers’ portfolios,

and 237 customers’ portfolios.

Table 1 shows sample characteristics such as distributions of downgrades by issuer, industry, and the

number of rivals, suppliers and customers for downgraded firms in this study. In panel A, average down-

grade events per issuer is 3.58, the median is 2.5, and the standard deviation is 3.39. The maximum num-

ber of downgrade events for one issuer is 19 and the minimum number of downgrade events for one

issuer is 1. Total number of debt issuers is 100. Panel B displays the distribution of downgrade events by

industry according to one digit SIC code of Compustat. Downgrades in manufacturing (1-digit SIC code:

J.-H. Chang et al. / Finance Research Letters 15 (2015) 232–238 235

2

w

p

fi

h

3

a

w

S

T

w

c

m

a

A

e

w

i

w

r

I

t

r

r

4

S

c

s

4

b

T

d

a

r

p

g

m

r

d

C

and 3) industries constitute more than half of downgrade events. About 13% of the sample are from

holesale and retail trade (1-digit SIC code: 5) industry.

Panel C in Table 1 lists the distributions of industry rival portfolios, supplier portfolios, and customer

ortfolios. Rivals are defined as having the same three-digit SIC code of Compustat as the downgraded

rm. On average, rival portfolio has 8.46 firms, supplier portfolio has 6.65 firms, and customer portfolio

as 1.65 firms. Downgraded firms have more suppliers than customers in our sample.

. Methodology

To examine changes in credit risk of industry rivals, suppliers and customers around credit events, we

pply the standard event study approach to the CDS spread of related firms. For industry competitors,

e construct an industry portfolio as an equally weighted portfolio of firms which have the same 3-digit

IC code as the event firm, continuous daily CDS spread and stock return data around the event window.

hen, we calculate industry cumulative CDS spread changes (CSCs) for a time interval based on the event

indows. Similarly, we follow Hertzel et al. (2008) to form portfolios of suppliers and customers and

ompute their CSCs.

We apply market-index-adjusted returns model to estimate abnormal changes of CDS as well. Since no

arket CDS index is available for the whole period of this study, we construct CDS indices following Jorion

nd Zhang (2007). All CDS samples are classified into five categories according to their credit ratings (1:

AA and AA ratings, 2: A rating, 3: BBB rating, 4: BB rating, 5: B and below ratings). Such classification

nables each category with enough samples to calculate market indices. To avoid extreme values problem,

e build the CDS indices using the median CDS spreads instead of the mean values. With constructed CDS

ndices, we calculate cumulative abnormal CDS spread changes (CASCs) as CASCi(t1, t2) = ASit2 − ASit1 , here ASit = Sit − Irt , Sit is CDS spread for reference entity i at day t and Irt refers to CDS index spread for

ating r at day t following Jorion and Zhang (2007).

We examine CSCs and CASCs of suppliers and customers during the period of rating downgrade events.

n addition, we divide the full sample into two subsamples based on whether the CSCs or CASCs to rivals to

he downgraded firm are greater than (or equal to) zero or less than zero. Then conditioning on rivals’ CDS

eactions, we again investigate suppliers’ and customers’ CSCs and CASCs when firms are downgraded by

ating agencies.

. Results

We illustrate bond rating downgrade effects for rivals, suppliers and customers in full sample in

ection 4.1. Section 4.2 demonstrates credit spread changes of suppliers and customers depending on

redit spread changes of intra-industry rivals. In Section 4.3, we further divide samples into several sub-

amples and explain the distinctions between these subsamples.

.1. Information transferring effect of bond rating downgrades

Credit risk reactions of industry rivals, suppliers and customers of downgraded issuers are measured

y cumulative credit spread changes (CSC) and cumulative abnormal spread changes (CASC). Panel A in

able 2 displays that industry portfolios have significant increases in CSC around 1-day and 5-day of the

owngrade events. The fractions of increasing CSC on the event day and around 1- or 5-day windows are

ll above 50%. However, CASCs (adjusted for CDS index spread for corresponding ratings) have positive

eaction but they are not significant in the result. Even though, we still have more than 50% of industry

ortfolios have positive CASCs on the event day and around 1 or 5 day windows.

Supplier portfolios also have significantly increasing CSCs around 1- and 5-day of the rating down-

rades (Panel B). Average CSC is 3.68 bps for 1-day window and 9.82 bps for 5-day window. In addition,

ore than half of supplier portfolios suffer from increases of credit risk around the announcement of bond

ating downgrades. We find no evidence that customer portfolios have significant reactions around the

owngrade events. Average CSC in 5-day window even shows a negative value and the fraction of positive

SC is less than 50%.

236 J.-H. Chang et al. / Finance Research Letters 15 (2015) 232–238

Table 2

Effect of downgrades on CDS spreads of rivals, suppliers and customers.

[t1 , t2 ] CSC CASC

Mean t-test % (>0) Sign rank Mean t-test % (>0) Sign rank

(bps) p-value p-value (bps) p-value p-value

Panel A: industry portfolios (rivals) N = 358 [0,0] 4.16 0.13 64.53 0.00 3.80 0.17 53.91 0.01

[−1,1] 7.75 0.02 59.50 0.00 3.10 0.49 58.38 0.00 [−5,5] 18.15 0.01 58.10 0.00 7.60 0.25 56.15 0.01 Panel B: suppliers portfolios N = 155 [0,0] 1.32 0.21 63.23 0.03 −0.30 0.85 53.55 0.23 [−1,1] 3.68 0.08 56.13 0.10 0.83 0.72 57.42 0.23 [−5,5] 9.82 0.02 54.19 0.03 4.06 0.25 60.00 0.06 Panel C: customers portfolios N = 237 [0,0] 1.45 0.73 50.63 0.83 1.80 0.67 48.95 0.48

[−1,1] 0.81 0.94 50.63 0.73 1.20 0.91 49.37 0.46 [−5,5] −5.10 0.71 47.68 0.58 −7.30 0.56 45.57 0.17

Note: % (>0) denotes the percentage of observations with positive or zero values. The null hypothesis under the t-test is

mean-CSC (or CASC) = 0 and under the Wilcoxon sign rank test is median-CSC (or CASC) = 0.

To sum up, bond rating downgrades seem to have contagion effects within the industry and along the

suppliers of downgraded firms. Credit risk, measured by credit spreads of CDS, increases in industry and

supplier portfolios when bond rating downgrades are announced. However, we do not find information

transferring effects for customer portfolios in full sample.

4.2. Credit changes of suppliers and customers conditioning on rival portfolio reaction

To clearly identify contagion and competitive effects in rating downgrades, we separate full samples

into several subsamples. First, we analyze how vertical supply chain effects interact with horizontal con-

tagion and competitive effects in rating downgrades following Hertzel et al. (2008). We define horizontal

contagion effects as the average cumulative credit spread change (CSC) in 5-day window of the industry

portfolio is positive or zero values. Second, the contagion and competitive effects can depend on the origi-

nal bond rating of the downgraded firm (Jorion and Zhang, 2010) and hence we divide our sample into two

groups according to the original bond rating is investment-grade or not. Lastly, we identify the sample as

industry downgrade if other firms in the same industry as downgraded firms also suffer from downgrades

at the same time. These events can be resulted from clustering downgrades by rating agencies and hence

they are required to be examined separately.

In this section, we exhibit the first subsample test which examines credit changes of suppliers and

customers conditioning on the reaction of rival industry. Comparing with the result in full sample, we find

that CSC of suppliers reacts positively only when industry rivals have positive CSC (Panel A in Table 3). In

this horizontal contagion subsample, suppliers have 16.38 bps increase in 5-day window. Alternatively,

CSC of suppliers even decrease (not significant) conditioning on horizontal competitive effects (CSCs of

rivals are negative). Therefore, vertical contagion effects of suppliers happen only on horizontal contagion

condition.

Interestingly, we did not find significant effects of customers in full sample but we find vertical com-

petitive effects of customers conditioning on horizontal competitive situation. The average CSC of cus-

tomers in 5-day window drops 42.2 bps when the rivals have negative CSC. This means customers have

lower credit risk if rivals benefit from downgrades.

4.3. Reaction of suppliers and customers in subsamples

Panel B in Table 3 shows credit changes of suppliers and customers when the downgraded firms are

investment-grade or speculative-grade. We find that suppliers’ average CSC is 8.54 bps when the down-

graded firms belong to investment grade. However, when the downgraded firms are speculative-grade,

J.-H. Chang et al. / Finance Research Letters 15 (2015) 232–238 237

Table 3

Reaction of suppliers and customers CDS spreads in subsamples.

Panel A: suppliers and customers reaction conditioning on rivals reaction

Suppliers Customers

Rival CSC>0 Rival CSC<0 Rival CSC>0 Rival CSC<0

CSC (bps) 16.38 −0.04 22.45 −42.19 t-test p-value (0.00) (1.00) (0.21) (0.05)

%(>0) 68.82 32.26 57.35 34.65

Sign rank p-value (0.00) (0.01) (0.01) (0.00)

Obs. 93 62 136 101

Panel B: suppliers and customers reactions for downgraded firms belong to investment-grade

Investment grade Speculative grade Investment grade Speculative grade

CSC (bps) 8.54 13.75 −17.43 18.65 t-test p-value (0.04) (0.19) (0.20) (0.54)

%(>0) 51.28 63.16 46.79 49.38

Sign rank p-value (0.09) (0.29) (0.73) (0.71)

Obs. 117 38 156 81

Panel C: suppliers and customers reactions for industry downgrades

Industry downgrade Firm downgrade Industry downgrade Firm downgrade

CSC (bps) 20.59 5.94 −12.95 −3.41 t-test p-value (0.01) (0.21) (0.52) (0.83)

%(>0) 58.54 52.63 45.24 48.21

Sign rank p-value (0.04) (0.26) (0.83) (0.61)

Obs. 41 114 42 195

Note: % (>0) denotes the percentage of observations with positive or zero values. The null hypothesis under the

t-test is mean-CSC = 0 and under the Wilcoxon sign rank test is median-CSC = 0.

t

t

t

o

5

i

i

s

1

i

n

5

t

d

c

e

r

c

a

he average CSC is 13.75 bps but not significant. Thus, suppliers suffer from credit contagion only when

he downgraded firms are investment-grade companies.

For customers, average CSC is negative when the downgraded firms are investment-grade but is posi-

ive when they are speculative-grade. However, both of them are insignificant. Additionally, the fractions

f positive or zero CSC for customers in investment-grade and speculative-grade subsamples are less than

0%. Therefore, we find no evidence that shows customers react differently when downgraded firms are

nvestment-grade or not.

When more than one firm downgraded at the same time in one industry (i.e., industry downgrades),

t is expected that the effect will be distinct from the effect of firm-specific downgrades. We find that

uppliers suffer from 20.59 bps CSCs when downgrades belong to industry factors and it is significant at

% significance level. However, idiosyncratic factor downgrades only cause suppliers 5.94 bps CSCs and it

s insignificant. We do not find significant difference of CSCs for customers neither in industry downgrades

or in firm-specific downgrades.

. Conclusion

In this study, we examine information effects of bond rating downgrades in intra-industries and along

he supply chains. We find that suppliers suffer from vertical credit contagion effects when the industry of

owngraded firms also suffers from credit risk increase (horizontal credit contagion effects). In addition,

ustomers experience vertical credit competitive effects in occurrence of horizontal credit competitive

ffects. These results help us comprehensively understand credit information transferring effects among

elated companies and clustering reactions in CDS markets. It is particularly useful when applying to

onstruct credit portfolios and to manage credit risk for financial industries.

This research is subjected to unavailable data of suppliers and customers for some downgraded firms

nd lacks of CDS spreads for suppliers and customers in our research period. Additionally, several research

238 J.-H. Chang et al. / Finance Research Letters 15 (2015) 232–238

directions are worth of further studies in the future. For example, we can study credit spread changes in

the debt market for suppliers and customers in these credit events. It is also interesting to study credit

contagion and competitive effects when bonds are upgraded.

Acknowledgment

We are grateful to Ministry of Science and Technology Taiwan (Grant nos. MOST 103-2410-H-468-005

and MOST 104-2410-H-468-010) for financial supports.

Reference

Alemany, A., Ballester, L., González-Urteaga, A., 2015. Volatility spillovers in the European bank CDS market. Financ. Res. Lett. 13,

137–147. Chang, J.-R., Hung, M.-W., Tsai, F.-T., 2012. Cross-market hedging strategies for credit default swaps under a Markov regime-switching

framework. J. Fixed Income 22, 44–56.

Delis, M.D., Mylonidis, N., 2011. The chicken or the egg? A note on the dynamic interrelation between government bond spreads and credit default swaps. Financ. Res. Lett. 8, 163–170.

Fabozzi, F.J., Cheng, X., Chen, R.-R., 2007. Exploring the components of credit risk in credit default swaps. Financ. Res. Lett. 4, 10–18. Goh, J.C., Ederington, L.H., 1993. Is a bond rating downgrade bad-news, good-news, or no news for stockholders. J. Financ. 48, 2001–

2008. Hertzel, M., Li, Z., Officer, M., Rodgers, K., 2008. Inter-firm linkages and the wealth effects of financial distress along the supply chain.

J. Financ. Econ. 87, 374–387.

Jorion, P., Liu, Z., Shi, C., 2005. Informational effects of regulation FD: evidence from rating agencies. J. Financ. Econ. 76, 309–330. Jorion, P., Zhang, G., 2007. Good and bad credit contagion: evidence from credit default swaps. J. Financ. Econ. 84, 860–883.

Jorion, P., Zhang, G., 2009. Credit contagion from counterparty risk. J. Financ. 64, 2053–2087. Jorion, P., Zhang, G., 2010. Information transfer effects of bond rating downgrades. Financ. Rev. 45, 683–706.

Lang, L.H.P., Stulz, R., 1992. Contagion and competitive intra-industry effects of bankruptcy announcements : an empirical analysis. J. Financ. Econ. 32, 45–60.

Odders-White, E.R., Ready, M.J., 2006. Credit ratings and stock liquidity. Rev. Financ. Stud. 19, 119–157.