Does the source of debt financing affect default risk?

• Author: Chih‐Wei Wang,Juan Ignacio Peña,Wan‐Chien Chiu
• Published date: 01 July 2018
• DOI: http://doi.org/10.1016/j.rfe.2017.03.006
• Date: 01 July 2018

ORIGINAL ARTICLE

Does the source of debt financing affect default risk?

Wan-Chien Chiu

1

|

Chih-Wei Wang

2

|

Juan Ignacio Pe~

na

3

1

Department of Quantitative Finance,

National Tsing Hua University, Taiwan

2

Department of Finance, National Sun

Yat-sen University, 70, Lianhai Rd.,

Kaohsiung 804, Taiwan

3

Department of Business Administration,

Universidad Carlos III de Madrid, Calle

Madrid 126, 28903 Getafe, Madrid, Spain

Correspondence

Chih-Wei Wang, Department of Finance,

National Sun Yat-sen University, 70,

Lianhai Rd., Kaohsiung 804, Taiwan.

Email: wanchien418@gmail.com,

chwang@mail.nsysu.edu.tw,

ypenya@eco.uc3m.es

Abstract

We examine whether the source of debt financing is important for assessments of

firms’default risk. This study reveals that during the 2007–2010 financial crisis,

firms that depend mainly on financing from banks suffer higher increases in

default risk than do firms with no such dependence. Conversely, firms that rely

solely on financing from public debt markets do not experience significant

increases in default risk. These findings suggest that the bank supply shock theory

explains the transmission of financial shocks to the real economy. Finally, firms

that depend on bank financing cannot offset the adverse impacts of bank lending

shocks by substituting bank loans with publicly traded debt.

JEL CLASSIFICATION

G01, G18, G32

KEYWORDS

Bank loans, Debt financing, Default risk, Public debt markets, Supply shock

1

|

INTRODUCTION

Debt financing literature indicates that the source of debts has impacts on firms’economic situations, investments (Cha va

& Purnanandam, 2011), and capital structure (Faulkender & Petersen, 2006). With a somewhat different approach, we focus

on the link between the source of debts and firms’default risk, which explains the manifestation of financial shocks as

reduced real investments.

1

Understanding what drives default risk during credit crunches thus is important for efforts to sta-

bilize the real economy. Our evidence suggests that the debt financing source strongly influences firms’default risk during

the 2007–2010 financial crisis.

How does the debt financing source affect the default risk of firms? Credit supply shock theory asserts that firms raising

funds mainly from credit markets face challenges, because the crisis affects all credit channel s (Gorton, 2010). Bank supply

shock theory further states that, in response to shocks in banking systems, banks choose not to renew loans and refrain

from issuing new loans. The impaired bank financing channel then generates stronger adverse impacts on firms that rely

more on financing from these banks (Ivashina & Scharfstein, 2010). Furthermore, bank-dependent firms with access to pub-

lic debt markets seemingly substitute their bank loans with publicly traded debts, which could mitigate the adverse impact

due to bank lending shocks. We find strong evidence in support of the bank supply shock theory, in that the default risk of

firms that depend on bank debt increases significantly more than the default risk of similar firms without such dependence.

However, we cannot affirm credit supply shock theory, because we find that the default risk of firms that rely solely on

public debts does not change significantly. Nor do substitution benefits arise; bank-dependent firms, regardless of their abil-

ity to access public debt markets, experience similar increases in their default risk.

The empirical findings reflect an appropriate selection of all listed non-financial firms in the U.S. market from 2006 to

2010, which produces a sample of 3169 unique firms and 113,409 firm-month observations. We measure default risk with

First published online by Elsevier on behalf of The University of New Orleans, 30 March, 2017, https://doi.org/10.1016/j.rfe.2017.03.006

Received: 20 July 2016

|

Revised: 16 March 2017

|

Accepted: 27 March 2017

DOI: 10.1016/j.rfe.2017.03.006

232

|

©2017 The University of New Orleans wileyonlinelibrary.com/journal/rfe Rev Financ Econ. 2018;36:232–251.

an indicator of the distance-to-default and analyze its time-series changes over the crisis period (relative to its level during

the pre-crisis period).

2

Furthermore, we adopt a difference-in-differences methodology, such that we compare the cross-sec-

tional heterogeneity of time-series changes in default risk across firms that rely on different debt financing source s but that

exhibit similar risk characteristics before the 2007–2010 crisis. This robust method reduces the concern that some unobserv-

able risk factors may drive the results. In addition, we focus on results during the early stage of the financial crisis. Because

this period is less affected by the demand-side effect, we provide more plausible results related to the supply shock effect.

3

Default risk increases more for firms that rely on banks for financing than for other types of credit-dependent or non-

credit-dependent firms. This evidence is not consistent with the notion that an overall credit crunch is the main channel of

transmission for the effects of the financial crisis from the financial sector to the real economy, but it does support the bank

supply shock theory. Risky firms are more exposed to supply shocks (Lemmon & Roberts, 2010), so if default risk

increases more for firms with low credit quality than for high credit quality firms, the supply shock effect receives further

support—as we show with this article.

We further examine whether the ability to substitute bank debts with public debts is beneficial for reducing the bank-

dependent firms’default risk during crises. For this test, we select only bank-dependent firms; then compare the distance-

to-default of borrowers that can access public debt markets versus the distance-to-default of borrowers that are unable to do

so. Our empirical evidence indicates that the substitution effect is less likely, because we find no economically or statisti-

cally important differences.

In our baseline analysis, we classify bank-dependent firms according to their repeated borrowing history from the

same lead bank, as obtained from the DealScan LPC database. For robustness, we also consider another identification

strategy in which we use the ratio of a firm’s bank debts to its total assets before the crisis. Newly identified bank-

dependent firms continue to experience a significantly greater increase in their default risk. We also perform a battery of

robustness tests to reduce concerns related to the bank loan database, matching method, debt maturity structure, or

whether firm value and asset volatility drive increased default risk. The results provide further support for bank supply

shock theory.

We accordingly make several contributions to prior literature. First, the extent to which distress in the financial sector

spills over to default risk in the real economy has remained largely unclear (see Chiu, Pe~

na, & Wang, 2015). Despite an

abundance of studies of the 2007–2010 crisis, ours is one of a handful to link shocks in the financial system to corporate

default risks empirically. Furthermore, we extend bank supply shock theory, often tested in emerging markets (e.g., Khwaja

& Mian, 2008; Schnabl, 2012), by focusing on the U.S. market and considering not just firms that borrow from banks but

all non-financial firms. Moreover, we complement substitution effect literature by directly examining whether accessing

public debt markets helps bank-dependent firms mitigate their default risks during weak economic times; previous studies

focus on select factors, such as firms’net investments, net debt issuances, or equity valuation losses (Chava & Purnanan-

dam, 2011; Lemmon & Roberts, 2010). Finally, our study evidence regarding the substitution effect on the default risk of

firms complements Santos and Winton’s (2008) assertion that altered risk for bank-dependent firms during recessions is not

significantly different from the changes in this risk for firms with access to public bond markets. We provide new evidence

in support of this prediction.

The remainder of this paper is organized as follows: Section 2 describes the data, default risk measure, and identification

strategy. Section 3 contains the main empirical results. In Sections 4 and 5 we offer additional robustness tests to support

the main findings. We conclude in Section 6 with a summary of the results and some further research suggestions.

2

|

DATA, DEFAULT RISK MEASURE, AND IDENTIFICATION STRATEGY

2.1

|

Sample construction

We examine all publicly listed, non-financial firms in the U.S. market from 2006 to 2010, according to a set of data sample

selection rules. First, we include only leveraged firms, which have obligations to pay for debt and thus could default. We

remove firms with zero debt in the second quarter of 2006 (before the crisis). Second, we exclude financial firms (standard

industrial classification [SIC] codes 6000–6999), utility companies (SIC codes 4910 and 4940), and firms in the public sec-

tor (SIC codes 9000–9999). Third, we choose firms with available daily equity prices and quarterly balance sheet informa-

tion in the Center for Research in Security Prices (CRSP) and Compustat databases. We lag all accounting information by

three months, to reflect reporting delays, and substitute missing accounting data with the most recent prior observations.

Thus we can obtain available information on firm risk characteristics, which we use to build a matched sample. The final

sample consists of 3169 unique firms and 113,409 firm-month observations.

CHIU ET AL.

|

233