Exploring short‐ and long‐run links from bank competition to risk

• Author: E. Philip Davis,Dilruba Karim
• Published date: 01 June 2019
• Date: 01 June 2019
• DOI: http://doi.org/10.1111/eufm.12176

D

462 © 2018 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/eufm Eur Financ Manag. 2019;25:462–488.

DOI: 10.1111/eufm.12176

ORIGINAL ARTICLE

Exploring short- and long-run links from bank

competition to risk

E. Philip Davis

1,2

|

Dilruba Karim

1

1

Department of Economics and Finance,

Brunel University, Uxbridge, Middlesex,

UB8 3PH, United Kingdom

Emails: philip.davis@brunel.ac.uk;

p.davis@niesr.ac.uk;

e_philip_davis@msn.com;

dilruba.karim@brunel.ac.uk

2

National Institute of Economic and

Social Research, 2 Dean Trench Street,

London SW1P 3HE, United Kingdom

Funding information

This paper was prepared under ESRC

Project No. ES/K008056/1, titled ‘The

Future of Banking’

Abstract

The current literature offers diverse findings on the bank

competition-risk relationship. We seek to advance under-

standing by looking at both short- and long-run relation-

ships for banks from 27 EU countries, using a six-year

period before and since 2007 and employing both the H-

statistic and the Lerner index as measures of competition.

We thus highlight further nuances in the competition–risk

relationship that are absent from the current literature. Both

measures have a positive short-run relationship with risk,

while long-run effects differ. Underlying this, the competi-

tion measures differ in their relationship to the volatility of

profits, with important policy implications.

KEYWORDS

bank competition, EU banking markets, financial stability, Lerner

index, Panzar–Rosse H-statistic, Z-score

JEL CLASSIFICATION

G21, G28

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction

in any medium, provided the original work is properly cited.

© 2018 The Authors. European Financial Management Published by John Wiley & Sons Ltd

The authors thank an anonymous referee, John Doukas (the Editor), Angus Armstrong, Ray Barrell, Thorsten Beck, Jaap

Bikker, Jerry Caprio, Dawn Holland, Iana Liadze, Corrado Macchierelli, Phil Molyneux, and Pawel Paluchowski for

assistance and advice. All errors remain our responsibility. This paper was prepared under ESRC Project No. ES/K008056/

1, titled ‘The Future of Banking’.

1

T

(

d

o

b

t

r

c

o

c

h

b

a

U

o

s

t

t

a

t

t

w

p

K

o

l

f

l

c

a

c

i

(

s

d

p

r

a

d

d

L

T

DAVIS AND KARIM

|

3

463

8.

n

1

|

INTRODUCTION

The subject of bank competition and risk has returned to the fore with the global financial crisis

(GFC) in 2008–2009, with a popular view being that competition between financial institutions

during the preceding boom was an important feature underlying the crisis (e.g., the majority view

of the Financial Crisis Inquiry Commission, 2011). This, in turn, would imply that the benefits of

banking competition for economic growth and efficiency need to be considered in the balance. On

the other hand, an extensive literature, generally estimated using pre-crisis data, finds diverse

results for the relationship between competition and risk. This follows, on the one hand, the so-

called franchise value or competition–fragility approach −that more competition reduces the value

of a banking licence and thus induces firms to take more risk −and, on the other hand, the

competition–stability view, that, with low levels of competition, banks may charge excessively

high rates of interest on loans and hence generate adverse selection and moral hazard on their loan

books. Both types of results have been found in the empirical literature. An emerging set of studies

also suggests that both high and low levels of competition may be adverse for risk, that is, there is a

U-shaped relationship, and that structural and regulatory features could affect country-level trade-

offs.

Further ambiguity was introduced by the common use of the concentration of a banking

system as a proxy for competition in earlier work in this area (e.g., Beck et al., 2006). However,

the theory of contestable markets suggests a concentrated system may be highly competitive if

there is sufficient potential competition from outside, as may be permitted by regulations

allowing new entry, as in the European Single Market. The empirical literature mostly covers

the period up to 2007, thus leaving open the interpretation of the post-crisis world, in which

there are a diminishing number of banks, extensive government intervention, and −many

would argue −less competition. The literature also suggests untested assumptions behind

proposals for enforced structural change in banking, such as the Vickers proposal in the United

Kingdom.

In this paper, we relate indices of banking market competition for the United Kingdom and

other European Union (EU) countries to banking risk. Our aim is to test for dynamic as well as

long-run links from competition to risk, before as well as after the GFC, thus highlighting

further nuances in the relation of competition to stability that have not been emphasized in the

literature to date. First, we use the Panzar–Rosse H-statistic to assess the changing nature of

competition in individual markets over time. Among earlier studies using this approach is an

analysis of the competition in the major Economic and Monetary Union (EMU) countries

compared to the United States just prior to the EMU (De Bandt and Davis, 2000). Second, we

investigate alternative approaches to measuring bank competition by using Iwata's Lerner index

(Bikker, 2004).

This paper is structured as follo ws. In section 2, we briefly examine the theoretical literatur e and

summarize recent empirical wo rk on competition and risk in bankin g, highlighting the fact that th e

datasets used largely cover the boom period or before, with rela tively little work on the post -GFC

period. We note that empirica l as well as theoretical work o ffers diverse findings and we pr obe the

reasons why. We question, for exa mple, whether global dataset s are fully informative for poli cy in

advanced countries. In sec tion 3, we outline the data and methodology we use in the exercise, before

discussing our empirical res ults. We test the impact of each meas ure of competition on risk,

differentiating between st atic and dynamic aspects and between the results from the H-statisti c and

Lerner index and showing se veral robustness checks, be fore considering reasons f or differences.

The final section draws our c onclusions.