The interplay between quantitative easing, risk and competition: The case of Japanese banking

• Published date: 01 February 2018
• DOI: http://doi.org/10.1111/fmii.12092
• Date: 01 February 2018

DOI: 10.1111/fmii.12092

ORIGINAL ARTICLE

The interplay between quantitative easing, risk

and competition: The case of Japanese banking

Emmanuel C. Mamatzakis Anh N. Vu

Universityof Sussex, School of Business,

Managementand Economics, Falmer, Brighton,

BN19SL, United Kingdom

Correspondence

AnhN. Vu, School of Business, Management

andEconomics, University of Sussex, Falmer,

Brighton,BN1 9SL, United Kingdom.

Email:A.Vu@sussex.ac.uk

Abstract

The Japanese economy is infamous for the magnitude of bank non-

performing loans that have originated back in the 1990s, whereas

they are still causing controversies. Japan is also known for an

extended quantitative easing programme of unprecedented scale.

Yet the links between risk-taking activities, quantitative easing and

bank competition are largely unexplored. This paper employs, for

the first time, the Boone indicator to measure bank competition in

Japan to examine these underlying linkages. Given the scale of non-

performing loans, we explicitly measure bank risk-taking based on

a new data set of bankrupt and restructured loans. The dynamic

panel threshold and panel Vector Autoregression analyses show

that enhancing quantitative easing and competition would reduce

bankrupt and restructured loans, but it would negatively affect

financial stability.Given the recent adoption of negative rates in Jan-

uary 2016 by the Bank of Japan, our study provides new insights as

clearly there is a trade-off between quantitative easing and financial

stability beyond a certain threshold. Caution, therefore, regarding

further scaling up quantitative easing is warranted.

KEYWORDS

bank risk-taking, Boone indicator, dynamic panel threshold analysis,

Japan, Quantitative easing

JEL CLASSIFICATION

G21, C23, E52

1INTRODUCTION

The competition and bank risk-taking nexus has sparked heated debates (Beck, De Jonghe, & Schepens, 2013; Boyd

& De Nicolo, 2005; Jiménez, Lopez, & Saurina, 2013; Tabak,Gomes, & da Silva Medeiros, 2015). There is extensive

research that reveals the mixed evidence as both positive and negative relationships between bank competition and

c

2018 New YorkUniversity Salomon Center and Wiley Periodicals, Inc.

Financial Markets,Inst. & Inst. 2018;27:3–46. wileyonlinelibrary.com/journal/fmii 3

4MAMATZAKISAND VU

risk are reported (Berger, Klapper,& Turk-Ariss, 2009; Fiordelisi & Mare, 2014; Fu et al., 2014; Liu & Wilson, 2013).

One strand of the literature argues that there are benefits to be derived from enhanced competition as it promotes

efficiency and preventsbanks from taking excessive risk (Schaeck & Cihák, 2008; Stiroh & Strahan, 2003). On the other

hand, some (Fu, Lin, & Molyneux, 2014; Liu & Wilson, 2013) raiseconcerns due to uncertainties that could be brought

by increased competition through excessivebank risk-taking. Others argue that stiff competition results in loss of high

economic rents (e.g.lending opportunities and profit) associated with reduced competition, and this has an augmenting

effect on risk (Allen & Gale, 2004; Keeley,1990). Hence, whether higher competition destabilises the banking system

by accumulating bank risk remains yet to reach unanimity.

In this paper, we build on the existingliterature to investigate the relationship between bank competition and risk

for the Japanese banking industry that also experiences extensive quantitative easing. To this end, we bring into the

framework also quantitative easing. We further innovate byusing bank specific Boone indicators instead of its aggre-

gate value or the Lerner indexto capture competition.

The case of Japan is of interest as it has faced chronic problems with nonperforming loans, and is one of the first

economies that an extensiveand far-reaching program of quantitative easing has been initiated.1We tackle the former

factor in our measure of bank risk-taking by opting for a new data set, whilst we explore the impact of quantitative

easing through the bank risk channel. Given the significance of quantitative easing for Japan, it warrants examining

its impact at bank level. The emerging of quantitative easing as a monetary policy tool to achieve price stability has

raised concerns among academics and policymakers about its association with bank risk-taking (Chodorow-Reich,

2014; Claeys & Darvas, 2015). Low short-term interest rates prior to loan issuance result in banks grantingmore new

risky loan portfolios, distorting their credit supply to favour borrowers with worse credit histories, lower ex-ante

internal ratings, and weaker ex-post performance (Ioannidou, Ongena, & Peydró, 2015; Jiménez, Ongena, Peydró,

& Saurina, 2014). Less return from yields is another motive for financial institutions to accelerate their risk-taking

activities (Chodorow-Reich, 2014; Rajan, 2005). Banking surveys based on credit standards in the US and the UK, on

the contrary, do not suggest an excessiverisk-taking by banks as a result of the enforcement of quantitative easing

(Claeys & Darvas, 2015).

The Bank of Japan was the pioneer in empowering quantitative easing policy. Currently,there has been a strong

record of activeand aggressive quantitative easing since 2010. We are interested in investigating whether the warning

of heightened financial stability risks associated with this policy is supported by Japanese bank leveldata. We hypoth-

esise, based on the aforementioned literature, that quantitative easing and competition affect bank risk-taking. After

the acute phase of the banking crisis in Japan (1997-1999), the banking system underwent major reforms, bailout and

consolidation from 1999 to 2003. Their competition stance, hence, is expected to vary over time, also in light of the

global financial crisis 2007–2008. Between 2000 and 2015, quantitative easing was launched twice (during March

2001-March 2006 and from October 2010). This could be considered as a macroeconomic shock to bank competi-

tion due to the relaxation of economic conditions, which in turn may affect the competition-risk nexus. The degree of

competition in the banking industry,however, could also influence the quantitative easing-risk linkage (Altunbas, Gam-

bacortab, & Marques-Ibanezc, 2014). Therefore, we control for the effect of quantitative easing when measuring the

competition-risk relationship, and vice versa. We also explore in depth the underlying causality among quantitative

easing, competition and risk.

Thereby, our study contributes to the literature in the following ways. First, whereas the current literature has

mostly used the bank-specific Lerner index or aggregate Boone indicator as a proxy for competition (Liu & Wilson,

2013; Schaeck & Cihák, 2014), we estimate bank level Boone indicators. We use local regression techniques to calcu-

late the Boone indicator for each bank-year observation. Second, we opt for an original data set to capture risk that

has been overseen by the literatureto date. Bank risk-taking, our primary focus, is represented by bankrupt loans and

restructured loans. Data on bankrupta nd restructured loansare available for Japanese commercial banks and have not

beenused extensively in the Japanese banking literature (Mamatzakis, Matousek, & Vu, 2015). We also use the classical

measure of bank default risk, Z-score, to enhance the robustness of our analyses. The use of bankrupt and restructured

loans at semi-annual data frequency allows an enriched information set in our modelling of competition and quanti-

tative easing.2Third, we employ a bank level proxy of quantitative easing that is the bank specific lending rate.The

MAMATZAKISAND VU 5

advantages of this microeconomic measure lie on the absence of aggregation bias and the ample set of information.

This bank-specific variable ensures its compatibility with the bank level Boone indicator and risks in our analyses. We

also conduct the analyses with two other proxiesfor quantitative easing: the 10-year Japanese government bond yield

and Bank of Japan total assets. Toexamine the risk-competition and risk-quantitative easing nexus,we employ dynamic

panel threshold analysis, where Generalised Methods of Moments type estimators are used to tackle the issues of

endogeneity (Kremer, Bick, & Nautz, 2013). This methodology allows us to examine whether these relationships are

stable over the observed period (financial years from 2000 to 2014) which embraces quite a few important events.

Theyare the final phase of the banking crisis (2000-2001), the restructuring period (2001-2003), the presence of quan-

titative easing (2001-2006 and from 2010), the global financial crisis (2007-2008),and the Tohoku earthquake (2011).

Fourth,we extend our analysis by using a panel vector autoregression (VAR)approach to address the underlying causal-

ity and the potential endogeneity among competition, quantitative easing and risk.

Our results show that competition and quantitative easing appear to undermine overall bank stability. However,

quantitative easing reduces problem loans. The results could entail the countervailing effects of quantitative easing

on bank risk-taking (Buch, Eickmeier,& Prieto, 2014; De Nicolò, Dell'Ariccia, Laeven, & Valencia, 2010). Regarding the

causality between the variables of interest, the panel VARanalysis suggests that quantitative easing causes bank risk

and bank competition.

The paper proceeds as follows. Section 2 briefly reviews the literature and associated hypotheses. Section 3

presents the methodologies. Section 4 introduces the data. Section 5 discusses the results. Finally,Section 6 concludes.

2RELATED LITERATURE AND HYPOTHESES

In this section, we establish our research hypotheses based on the literature regarding the competition-risk nexusand

the relationship between quantitative easing and bank risk. The two renowned hypotheses about the impact of com-

petition on financial stability, an important part of our main investigation objectives, have been well defined in the

literature as introduced in the following sections.

2.1 The competition – fragility hypothesis

The underlying theory of this hypothesis poses the view of uncertainty created by a competitive banking industry.

The rationale behind this is the threat of market share being reduced by the entry of newly established banks as well

as stronger competence of incumbent rivals. The rise in bank competition could be attributed to, e.g., consolidation,

deregulation, and technological advances (Berger & Mester,2003; Jeon, Olivero, & Wu, 2011; Keeley, 1990). The liber-

alisation of geographicrestriction and relaxation in unconventional banking activities have also fostered bank competi-

tion (Berger & Mester,2003). There exists evidence suggesting that when deregulation took place (e.g. in the US during

1970s–1980s), poor performers were more vulnerable due to the incompetence in keeping pace with their counter-

parts and potential entrants (Stiroh & Strahan, 2003).Deregulation also fuelled bank consolidation, resulting in a large

number of banks disappearing from the market due to mergers (Berger & Mester,2003). Note that sizeable consoli-

dation could, however,result in large banks exerting considerable market power (Yildirim & Mohanty,2010). In devel-

oping banking markets, beside the lift in entry barriers, technological development is another catalyst for heightened

competition brought by foreign bank entry (Jeon et al., 2011).

One of the main arguments for greater risks corresponding to increased competition is profit reduction. This rea-

sonably serves as a motive for bank managers to take excessiverisks to pursue business targets, to preserve market

shares, and eventually to protect market power.Of course, the notion of falling profitability could also raise concerns

among banks’ executivesand jeopardise their position. Consequently, they may have the incentive to stretch their risk

toleranceability. Keeley (1990) is among the studies laying the first bricks of the debate of an increasing levelof fragility

in association with intensified competition. The results lend support for the hypothesis to the extent that amplified