The impact of corporate fraud on director‐interlocked firms: Evidence from bank loans

• Published date: 01 January 2019
• Author: Frank M. Song,Tat‐kei Lai,Adrian C.H. Lei
• Date: 01 January 2019
• DOI: http://doi.org/10.1111/jbfa.12362

DOI: 10.1111/jbfa.12362

The impact of corporate fraud

on director-interlocked firms: Evidence from

bank loans

Ta t - ke i L a i 1Adrian C.H. Lei2Frank M. Song3

1Departmentof Economics & Quantitative

MethodsIÈSEG School of Management (LEM-

CNRS9221)

2Facultyof Business Administration University of

Macau

3Economicsand Management School Wuhan

University

Correspondence

Tat-keiLai,Department of Economics

&Quantitative Methods IÈSEG School

ofManagement (LEM-CNRS 9221).

Email:t.lai@ieseg.fr

JELClassification: G30, G32

Abstract

We examine the impact of corporate fraud committed by one firm

(the “fraudulent firm”) on other firms with interlocking directors (the

“interlocked firms”), focusing on the debtholder side. We argue that

the revelationof a fraudulent firm's fraud can damage the reputation

of the interlocked firms because corporate governance can propa-

gate via director interlocks. Empirically,we find that the interlocked

firms'cost of debt is higher and the loan covenants become stricter

after the fraud cases of the fraudulent firms are revealed. Consis-

tent with the corporate governance propagation explanation, our

results are weaker (stronger) for interlocked firms that have bet-

ter (worse) pre-event corporategovernance standards. Our findings

suggest that corporate fraud of fraudulent firms can affect other

firms through director-interlocks beyond shareholder value.

KEYWORDS

agency conflicts, corporate fraud, corporate governance, cost of

debt, director appointments, director interlocks, earnings manage-

ment, loan covenants, reputation, spillover

1INTRODUCTION

Regulators and the market penalize firms that are involved in fraudulent activities. Parties directly related to these

fraudulent firms, such as shareholders, managers, and debtholders, suffer from various kinds of losses (see, e.g.,

Chava, Huang, & Johnson, 2018; Graham, Li, & Qiu, 2008; Karpoff & Lott, 1993; Karpoff, Lee,& M artin, 2008a; Kar-

poff, Lee, & Martin, 2008b; Yuan& Zhang, 2015). Are parties not directly but closely related to the fraudulent firms

also affected? This is plausible since firms are usually interconnected so that corporate policies of one firm or events

affecting one firm likely affect other firms via such connections.1

In this paper, we examine the contagion effects of corporate fraud on other firms sharing common directors

with the fraudulent firms, focusing on the debtholder side. Why can corporate fraud of the fraudulent firms affect

director-interlocked firms? Although interlocked firms are not explicitly involved in the fraudulent activities and

1As anecdotal evidence, after the Enron scandal in 2001, there was strong opposition to the re-election of FrankSavage as the director of Lockheed Martin

because he also sat on the board of Enron. See “Lockheed Director is Under Fire Because of Role on Enron Board,” in Wall StreetJournal, March 25, 2002,

http://www.wsj.com/articles/SB1017012957165383000.

32 c

2018 John Wiley & Sons Ltd wileyonlinelibrary.com/journal/jbfa JBus Fin Acc. 2019;46:32–67.

LAI ET AL.33

FIGURE 1 Research question of this paper and findings of Fich and Shivdasani (2007) and Graham et al. (2008)

[Colour figure can be viewed at wileyonlinelibrary.com]

therefore not legally liable, they can also be affected because corporate practices and corporate governance can

propagate across firms through common directors (e.g., Bizjak, Lemmon, & Whitby, 2009; Bouwman, 2011; Cai,

Dhaliwal, Kim, & Pan, 2014; Chiu, Teoh, & Tian, 2013; Kang & Tan, 2008). Therefore, corporate fraud should reveal

new information to the shareholders and creditors about the interlocked firms'corporategovernance standards. Why

do we focus on the debtholder side? Fich and Shivdasani (2007) show that firms interlocked with other fraudulent

firms experience significant negative equity market reactions in short event windows due to the reputation effect.2

Since corporate fraud of the fraudulent firms provides new information to the shareholders and debtholders of the

interlocked firms but their interests are not necessarily aligned (Smith & Warner,1979), they may respond differently

to the fraud events. Therefore, it is also important to examinethe responses of interlocked firms'creditors in order to

fully understand the impact of corporate fraudon the interlocked firms.

Conceptually,lenders set the interest rates for a firm's loans based on lenders'perceptions about the firm'scharac-

teristics. Graham et al. (2008) find that, following the revelationof a fraud event, lenders update their beliefs about the

fraudulent firm and raisethe cost of debt. To the extent that governancestandards can propagate across firms through

interlocking directors, if lenders believe that, after the fraud events, the financial quality of the interlockedfirms has

been compromised or these firms'probability of committing corporate fraud increases, then they will adjust the loan

price or opt to renegotiate the loan contract terms. Figure 1 illustratesour research question relative to the findings of

Fich and Shivdasani (2007) and Graham et al. (2008).

Our main data come from the fraud database of Karpoff,Koester, Lee, and Martin (2017) (hereafter “KKLM”) which

identifies the fraud cases (the “events”), the directordata from the ISS RiskMetrics database, and the bank loan data

from DealScan (Chava & Roberts, 2008). In the baseline regression sample, there are 199 firms (interlocked with 70

fraudulent firms) with 3,759 loans over the period between 1990 and 2013. Wefind that the cost of debt of the inter-

locked firms increases after the revelation of corporate frauds by the fraudulent firms. These results are robust to

subsample regressions, alternative estimation methods, and a placebo test. In terms of economic significance, we find

that the post-event loan spreads of the interlocked firms increase byabout 14.8%. This increase is roughly equivalent

to a 10.8 basis-point increase in the loan spread or US$1.29 million in annual interest payments per loan; these magni-

tudes are about one quarter and 30% of the direct impact of the eventson the fraudulent firms in terms of the increase

in loan spreads and the increase in annual interest payments, respectively.We also find that the interlocked firms'loan

terms (measured by the loan covenant indexof Bradley & Roberts, 2015) become stricter after the events; this is con-

sistent with the idea that after the fraud events, creditors update their beliefs about the interlockedfirms and protect

themselves by more restrictive loan covenants. Overall,our results suggest that lenders of the interlocked firms also

respond to the fraud events.

If the interlocked firms are affected because of the propagation of corporate governance via interlocking direc-

tors (Bouwman, 2011), we should expect that the main results are weaker(stronger) for interlocked firms with better

2Morespecifically, Fich and Shivdasani (2007) find that, when financial fraud lawsuits are filed against the fraudulent firms, firms interlocked with fraudulent

firmsexperience significantly negative abnormal returns of −0.98%and −0.92%over the event windows [−1,0]and [0,0], respectively.

34 LAI ET AL.

(worse) pre-event corporate governance standards. Empirically,using the Corporate Governance Index of Gompers,

Ishii, and Metrick (2003) and the Entrenchment Index of Bebchuk et al. (2009), we find consistent evidence to sup-

port this hypothesis. In addition to corporate governance propagation, the literaturehas documented other potential

mechanisms through which the negative impacts of fraud may transfer from the fraudulent firms to other firms. For

instance, firms may suffer if theyare in the same industries as the fraudulent firms (Gleason, Jenkins, & Johnson, 2008;

Goldman, Peyer,& Stefanescu, 2012; Yuan & Zhang, 2015) or if they are located near the fraudulent firms (Giannetti &

Wang, 2016; Parsons et al., 2016). To ensure that our main results are not driven by these other factors, we estimate

two more sets of regressions. First, we exclude firms that are in the same industries (2-digit or 3-digit SIC) as those

of the fraudulent firms and re-estimate the regressions. Second, we excludefirms whose headquarters are geographi-

cally close (within 25 km or 50 km) to the headquarters of the fraudulentfirms and re-estimate the regressions. In both

tests, we find that the baseline results are unaffected. These results, taken together,show support for the “corporate

governance propagation”explanation.

What can the interlocked firms do to minimize the adverse consequences? We find that the interlockedfirms are

more likely to terminate director appointments after the eventsand the relationship is weaker for independent direc-

torsor audit committee members but is stronger for “busy” directors (who have three or more outside director appoint-

ments). Moreover, we find that the post-eventcost of debt is lower for those interlocked firms that terminate more

director appointments. These results seem to suggest that the interlocked firms may be able to mitigate the negative

consequences by keeping the higher-quality directors (independent directors and audit committee members) while

terminating the lower-quality directors (“busy” directors).

Finally,we examine whether the increased cost of debt of the interlocked firms is because these firms tend to manip-

ulate earnings after the events. Using severalearnings manipulation measures following Dechow, Sloan, and Sweeney

(1996), we find that the interlocked firms are less likely to manipulate earnings after the events. In other words, the

increased cost of debt after the events is unlikelyto be driven by earnings manipulation of the interlocked firms.

Our paper contributes to the literature related to corporate fraud and director interlocks. The corporate fraud

literature identifies various consequences of corporate fraud. Fich and Shivdasani (2007) find that the marketreacts

negativelyto the interlocked firms after the fraud events. Tothe best of our knowledge, we are not aware of any studies

examining the impact of corporate fraud on the creditors of the director-interlocked firms. Our findings fill the gap

in the literature and complement Fich and Shivdasani (2007) by showing that debtholders of the director-interlocked

firms also respond to the fraud events. Relative to Graham et al. (2008) who find a direct impact of fraud eventson

the costs of debt of the fraudulent firms, we also find a sizable impact of fraud eventson the costs of debt of the firms

interlocked with the fraudulent firms. Our results are consistent with the findings in the director interlocks literature

that corporate governance can propagate across firms through common directors (e.g., Bouwman, 2011). These

results are unlikely driven by the possible spillover via industry peers or geographical proximity. The main message

of our paper is that the corporate fraud of fraudulent firms can affect other firms through director interlocks beyond

shareholder value.

This paper is organized as follows. Section 2 reviewsthe literature and develops the hypotheses. Section 3 describes

the empirical specification and the data. Section 4 presents the main empirical results about the contagion effects of

fraud on the cost of debt and loan covenants of the interlockedfirms. In Section 5, we further examine how the inter-

locked firms may mitigate the adverse consequences through changing director appointments and whether their cost

of debt is related to broader measures of earnings manipulation. Finally,Section 6 concludes the paper.

2LITERATURE REVIEW AND HYPOTHESIS DEVELOPMENT

2.1 Consequences of corporate fraud on fraudulent firms

Various studies in the extantliterature have evaluated the effect of corporate fraud on the parties directly related to

the fraudulent firms. Karpoff and Lott (1993) find that the reputation cost of corporate fraud is large as firm value is