How Banking Competition Changed over Time

1

• How Banking Competition Changed over Time

• Jacob Bikker (DNB, UU) Laura Spierdijk (RUG)
• Eurobanking 2008, 18-21 May 2008
• Maribor, Slovenia

2
Introduction

• During the past years several developments have
  affected the banking sector. For example
• Liberalization deregulation
• The European economic and monetary union
• Consolidation
• The progress in information and communication
  technology
• How have these developments affected competition
  in the banking sector during the past 15 years?

3
Outline

• Given a multitude of major developments over time
  with respect to competition, we investigate
  whether and how banking competition has changed
  over time.
• 101 countries, including all major Western
  economies
• Special attention is paid to Europe (EU15,
  Eastern-Europe that joined EU in 2004) and
  Emerging countries

4
Approach

• Using the Panzar-Rosse model proposed by Rosse
  and Panzar (1977) and Panzar and Rosse (1987), we
  estimate H statistics as a measure of banking
  competition.
• We assess changes over time in three different
  ways
• explorative yearly estimates of competition (11
  major countries and 2 regions)
• gradual changes over time (101 countries)
• structural breaks in competition (11 major
  countries and 2 regions)

5
Data sample

• Bankscope data covering 18,467 banks in 101
  countries over 16 years, containing in total
  112,343 bank-year observations.

6
Yearly estimates (1)
7
Yearly estimates (2)
8
Recursive estimates (1)
9
Recursive estimates (2)
10
Parametric approach

• Adjusted P-R model, allowing competition to
  change gradually over time

11
Estimation results

• Significantly downward trend in competition 39
  countries.
• Significantly positive trend in competition 22
  countries.
• No significant changes over time 52 countries.
• Emerging markets average annual rise in H of
  0.67.
• Remaining countries average annual decline in H
  of 1.8.
• EU15 and Eastern Europe downward trend in H.

12
Robustness check

• Linear effect of time may be restrictive.
• Polynomial and cubic specifications
• Linear model captures the general trend H over
  time, but banking competition may develop in a
  more complex way than the linear model suggests.

13
Structural stability

• Is the H statistic constant over time?
• We do sup-Wald tests for structural stability
  with a single unknown breakpoint in the style of
  Andrews (1993, Econometrica).
• The test determines a breakpoint. By means of a
  parametric bootstrap (Dieboldt and Chen, 1996) we
  can check whether the break point reflects a
  significant break.
• Value of H statistic before and after the break.
• Robustness checks avg-Wald and exp-Wald tests.

14
Estimation results (1)
15
Estimation results (2)

• With exception of Italy, all countries considered
  that joined the EMU feature a significant
  structural break in either 2001 or 2002.
• This is some years after the formal establishment
  of the EMU in 1999 and the introduction of the
  virtual euro.
• For the non-EMU countries Denmark and the United
  Kingdom two out of three tests suggest that there
  is no significant break, whereas in Switzerland a
  significant break emerges in 1995.

16
Estimation results (3)

• For Italy we establish an early break in 1991.
• For Japan we find a break in 2003.
• For the United States we detect a break in 2001.
• Apart from Japan, competition in all 11 countries
  and two regions decreased over time.

17
Overall robustness check

• Same methods applied to balanced samples.
• Smaller samples but similar outcomes.

18
Conclusions (1)

• On average, the changes in competition over time
  in the 101 countries under consideration are
  small, but they turn out substantial for several
  countries and regions.
• The competitive climate in the euro area was
  subject to a major break a few years after the
  introduction of the euro, initiating a period
  with less competitiveness.
• Various Western economies have undergone a
  significant decline in banking competition during
  recent years.

19
Conclusions (2)

• The part of Eastern Europe that now belongs to
  the European Union has also experienced a
  significant decrease in competition during the
  past decade, although the actual change in the
  competitive climate in this area is relatively
  modest.
• By contrast, the banking industry in emerging
  markets has become more competitive during the
  last decade.

20
Conclusions (3)

• We attribute the predominantly downward trend in
  competition to increased bank size and the shift
  from traditional intermediation to off-balance
  sheet activities.

21
Thank you for your attention