Corporate governance and correlation in corporate defaults

• Author: Yang (Greg) Hou,Jayasuriya M.R. Fernando,Leon Li
• Date: 01 May 2020
• DOI: http://doi.org/10.1111/corg.12306
• Published date: 01 May 2020

ORIGINAL ARTICLE

Corporate governance and correlation in corporate defaults

Jayasuriya M.R. Fernando

1,2

| Leon Li

1

| Yang (Greg) Hou

3

1

Waikato Management School, University of

Waikato, Hamilton, New Zealand

2

Faculty of Commerce and Management

Studies, Department of Finance, University of

Kelaniya, Kelaniya, Sri Lanka

3

University of Waikato, Hamilton,

New Zealand

Correspondence

Leon Li, Waikato Management School,

University of Waikato, Hamilton,

New Zealand.

Email: leonli@waikato.ac.nz

Abstract

Research Question/Issue: This study examines the effect of weak corporate gover-

nance in terms of concentrated ownership, low board effectiveness, low financial

transparency and higher shareholder rights on default correlation when firms have

different credit qualities.

Research Findings/Insights: Using historical default data in the United States from

2000 to 2015, we find that the degree of default correlation increases disproportion-

ately for firms with concentrated ownership, low board effectiveness, low financial

transparency and disclosures, and higher shareholder rights. More importantly, the

effect of weak corporate governance on default correlation is high during a financial

crisis.

Theoretical/Academic Implications: This is one of the first studies testing the impact

of corporate governance on the correlation in corporate defaults. It indicates new

avenues of research for both corporate governance and credit risk management in

relation to why joint default probabilities vary among firms.

Practitioner/Policy Implications: Our results imply that good corporate governance

is essential for credit risk management because poor corporate governance may

increase individual default risk and create the domino effect of credit defaults. Practi-

tioners and policy makers should enhance control over poor governance practices to

reduce the probabilities of default. Moreover, the impact of corporate governance on

correlation in corporate defaults is more pronounced in financial crises and warrants

consideration from policy makers to take steps toward cushioning its effects.

KEYWORDS

corporate governance, credit quality, default correlation

1|INTRODUCTION

Evaluating the default correlation among credit portfolios is a crucial

process for bank credit risk management. Default correlation pertains

to the relationship between a firm's individual default probability and

the joint default probability among firms. Thus, the estimation of

default correlation depends not only on the individual firm's default

probability, but also on joint default probabilities among firms. For

example, in a crisis, default correlation arises as a result of events of

particular firm default and the collective default among firms.

Researchers have identified cyclical correlation, the contagion effect,

and learning from others as factors that cause joint default probabili-

ties over the last few decades (Das, Duffie, Kapadia, & Saita, 2007).

However, why joint default probabilities vary among firms is still

under investigation. In this article, we provide convincing reasons for

the variations, based on the results of an empirical investigation of

firms’corporate governance practices. Thus, we make theoretical con-

tributions by testing the impact of corporate governance on corporate

default correlation. In brief, we hypothesize that the degree of default

correlation increases disproportionately for firms with weak corporate

Received: 24 February 2019 Revised: 9 October 2019 Accepted: 10 October 2019

DOI: 10.1111/corg.12306

188 © 2020 John Wiley & Sons Ltd Corp Govern Int Rev. 2020;28:188–206.wileyonlinelibrary.com/journal/corg

governance. The dynamic nature of default correlations due to corpo-

rate governance implies that weak corporate governance increases

individual default risk and may also accompany a disproportionate

increase in the credit risk of a portfolio.

The issue of correlation in corporate defaults has generated con-

siderable interest and research has begun to focus on methodology

and the examination of factors that cause defaults. Default correla-

tions can be estimated by any one of three methods. The first method

is to estimate default correlations based on asset correlations

(e.g., Das et al., 2007). The second method is to use the credit default

swap (CDS) or bond spread data (e.g., Jorion & Zhang, 2007). The third

method proposes using a standard binomial approach to measure

default correlation based on realized historical default data

(e.g., Lucas, 1995). Notably, default correlations have not been satis-

factorily modelled. Lucas's method, which we adopt in this research,

takes advantage of model-free estimation techniques (see Li &

Chen, 2018).

Factors that cause correlation in corporate defaults have

attracted considerable research interest. Das et al. (2007) identify the

three main reasons for default clustering as cyclical correlation

(Duffie, 1998; Keenan, Hamilton, & Berthault, 2000; Duffie, Saita, &

Wang, 2007); contagion effect (Aharony & Swary, 1983; Giesecke,

2004; Lang & Stulz, 1992); and learning from defaults (Jarrow & Yu,

2001). A recent study by Li and Chen (2018) examines liquidity, sys-

tematic risk, and size as extensions of the three primary sources of

default correlation proposed by Das et al. (2007).

This article contributes by hypothesizing and examining the

impact of corporate governance on the correlation in corporate

defaults. Some extensive studies have been undertaken to test the

effect of corporate governance on individual corporate defaults (see

Daily & Dalton, 1994). We have been unable to find any studies that

systematically investigate the impact of corporate governance on cor-

relations in corporate defaults. Following Standard and Poor's (2002)

governance framework, we define four critical domains of corporate

governance: ownership structure and influence, board effectiveness,

financial transparency and disclosures, and shareholder rights. Accord-

ingly, we develop four research hypotheses regarding the impact of

corporate governance on correlations in corporate defaults. Our

empirical results further indicate that default correlation is high for

firms with concentrated ownership, low board effectiveness, low

financial transparency and disclosures, and higher shareholder rights.

We follow the approach of Lucas (1995) by testing the impact of

corporate governance on correlations in corporate defaults using the

realised default data in the U.S. for the period between 2000 and

2015. In addition, following the studies of Lemmon and Lins (2003)

and Erkens, Hung, and Matos (2012), both of which indicate that the

effect of corporate governance on firm performance varies during a

financial crisis, we add a fifth hypothesis relating to defaults in both

crisis and non-crisis periods.

The results of this research can provide policy implications to

banks and regulatory authorities. As hypothesized in this study, the

degree of the domino effect of credit defaults increases dispropor-

tionately for firms with weak corporate governance, and the

phenomena are more pronounced during crisis periods. Regulators

should consider these findings and develop a viable regulatory capital

framework for credit risk management to mitigate the potential con-

sequences of underestimating default clustering due to poor corpo-

rate governance. Our empirical results imply that poor corporate

governance might cause an increase in individual default risk and also

exaggerate the domino effect of credit defaults. The increase in

default correlations will, in turn, reveal an increase in portfolio credit

risk. If the impact of corporate governance on default correlations is

not given consideration, the benefit of reductions in risk stemming

from credit portfolio diversification is likely to be overestimated for

firms with weak corporate governance. Firms with weak corporate

governance are associated with higher default correlations and, hence,

are less effective in risk reduction if they are included in credit portfo-

lio diversification efforts. Our findings support the argument that reg-

ulators should further adjust capital requirements for banks which

provide loans to weak governance firms in crisis periods.

The rest of the article is organized as follows. Section 2 reviews

the literature on sources of default correlation and modelling for

default correlations before developing the research hypotheses.

Section 3 describes the sample and research design. Section 4 pre-

sents the results of the empirical findings. Section 5 presents the

robustness tests undertaken in this study. Section 6 offers a conclu-

sion and considers future directions for research.

2|LITERATURE REVIEW AND

HYPOTHESES DEVELOPMENT

2.1 |Sources of default correlation

There are three main reasons for default correlation: common factors

or cyclical correlation, the contagion effect, and learning from others

(Das et al., 2007). Cyclical correlation arises as a result of a similar pat-

tern of correlated risk factors being present among firms. Common

economic factors include interest rates, inflation, GDP, business

cycles, and stock market performances. Duffie (1998) explains that

the aggregate default rates are correlated because of general interest

rate movements. Further, Duffie et al. (2007) find that personal

income growth and term structure levels also affect the changes of

default probabilities. De Servigny and Renault (2002) report that joint

default probabilities are higher in recession periods than in non-

recession periods. During a recession or a financial crisis, all businesses

are adversely affected by their sensitivity to the general economic

conditions. Li and Chen (2018) suggest that firms with a low beta,

which represents systematic risk, have high default correlation.

The contagion effect implies that one company's default induces

the default of another company.

1

For example, the default of a subsid-

iary company creates a default for its parent company. Accordingly,

the default clustering phenomena are invariably observed between

closely related companies with buyer-supplier relations. Initial studies

on the contagion effect have focused mainly on stock market informa-

tion (e.g., Lang & Stulz, 1992) while later studies have investigated

FERNANDO ET AL.189