Transcript Bank Supervisor Independence and the Health of Banking Systems

Bank Supervisor Independence
and the Health of Banking Systems
Evidence from OECD Countries
Steve Donzé
London School of Economics
[email protected]
Policy Background
• Just as central bank independence is now fairly
ubiquitous, financial supervisor independence has
gained momentum, especially in the last decade
• The need for independence first stressed in the
Basel Core Principles for banking supervision
(Basel Committee on Banking Supervision 1997)
• Financial supervisor independence is now a
standard policy prescription in the context of IMFled Financial Sector Assessment Programmes
2
Transmission mechanisms
• Financial supervisor independence enhances
policy credibility by ensuring financial
supervisors abide by legal requirements
– Government pressure generates timeinconsistent policies, thereby exacerbating
risk-taking bias/moral hazard
– Industry capture induces lax prudential
standards and/or enforcement
• Downside risk: independent supervisors may
not find optimal to act promptly (Kane 1990)
3
Motivation
• There is limited evidence that bank supervisor
independence matters for banking stability. E.g.,
Barth, Caprio, and Levine (2002); Das, Quintyn,
and Chenard (2004)
• This paper presents new indexes of bank
supervisor independence based on national
laws and assesses whether the formal adoption
of independent supervisory agencies results in
sounder banking systems
4
Legal Independence Indexes
• Set of statutory provisions enabling supervisory
agencies to resist political and industry pressures
• Legal bank supervisor independence from
government (LBSIG) summary index
– 4 sub-indexes (personnel, goal, policy and
budget independence)
• Legal bank supervisor independence from
industry (LBSII) summary index
– 2 sub-indexes (personnel and policy
independence)
5
MEX
NOR
JPN
AUS
DEU
SWE
KOR
IRL
ISL
ESP
CHE
CAN
HUN
DNK
NLD
ITA
NZL
TUR
LUX
GBR
FRA
SVK
PRT
USA
CZE
AUT
BEL
FIN
GRC
POL
LBSIG Summary Index
1.0
0.8
0.6
0.4
0.2
6
LBSIG Typology
Low personnel
independence
High personnel
independence
Low goal, policy and
budget independence
[Quadrant II]
Japan
Norway
Mexico
[Quadrant III]
Ireland
Korea
Spain
High goal, policy and
budget independence
[Quadrant I]
Austria
Denmark
Germany
[Quadrant IV]
Belgium
Finland
Poland
7
LBSIG Stylised Facts
• The most independent supervisors have not been
established in the more advanced economies
– Full independence is rare. Developed countries
discount independence by means of
accountability/political control (consistent with
Quintyn, Ramirez and Taylor 2006)
– Emerging countries implement good practices
in monetary and financial policies
• The higher bank supervisor independence, the
less unified financial sector supervision
– Politicians tend to fear ‘leviathan’ agencies
8
MEX
CHE
SVK
KOR
ITA
DEU
PRT
ISL
DNK
POL
LUX
ESP
NOR
NLD
GRC
HUN
FIN
TUR
SWE
NZL
IRL
CZE
BEL
AUT
AUS
JPN
FRA
GBR
USA
CAN
LBSII Summary Index
1.0
0.8
0.6
0.4
0.2
9
LBSII Stylised Facts
• Difference of means between central banks and
separate agencies is not significant anymore
• Common law countries outperform any other legal
origin sub-groups
– Due in part to the Anglo-Saxon bureaucratic
tradition, e.g., transparency in public life
• Bank-dominated financial systems are more likely
to have lesser independent supervisors
10
Empirical Relationships
• Do banking systems in countries with higher
independent supervisory authorities are
sounder?
• Do some dimensions of independence matter
more than others?
• The impact of legal independence is estimated
through cross-sectional regressions
– OLS with robust standard terms
11
Methodology
• Small sample size: OECD (30); OECD highincome (24)
• Basic equation
– Dependent variable
• Average bank financial strength ratings
from Moody’s and Fitch
– Variables of interest
• LBSIG summary/individual indexes
• LBSII summary/individual indexes
12
Methodology
• There is no single standard way to measure
banking sector soundness
– Financial statement data
– Market data
– Financial strength ratings
• Financial strength ratings
– Incorporate more information than any single
macro-prudential data
– Cross-country comparability
13
Methodology
• Control variables suggested by existing literature
– Bank characteristics (concentration; private
credit; foreign/state ownership
– Country characteristics
• Institutions (rule of law; veto players)
• Macroeconomic (inflation; GDP growth;
GDP per capita)
14
Main Regression Results
SAIR (dep. var.)
LBSIG
OECD (30)
(1)
(2)
.255*
(1.739)
LBSII
GDP per capita
(3)
High-income (24)
(4)
.244*
(1.683)
(7)
.257*
(1.753)
(8)
.367*
(1.850)
.113
(.738)
(9)
.420*
(1.932)
.252
(1.220)
.691*** .600*** .598***
(4.706) (3.918) (3.164)
.520***
(3.303)
GDP growth
.197
(1.400)
.370*
(1.714)
Inflation
-.177
(-.969)
-.139
(-.597)
Concentration
State ownership
N
R2
30
30
.093
(.634)
.161
(.747)
-.398**
(-2.522)
-.440*
(-2.054)
30
30
24
24
24
.454*** .405*** .512***
.570***
.135*
.063
.493*
15
Results
• LBSIG summary index is positively correlated with
Moody’s/Fitch ratings
– Not sensitive to control variables
– The impact of LBSIG is stronger at high levels
of rule of law, i.e., in high-income countries
– An increase in LBSIG summary index by one
std. dev. corresponds to 0.62 of a notch at Fitch
• By contrast, LBSII summary index, although
positively correlated, lacks robustness
– Capture is not necessarily destabilising
16
Sensitivity Tests
• Summary indexes were also calculated using
random weights instead of equal weights
– Results are robust with slight variations
• Using Moody’s or Fitch measures of banking
system soundness yields similar results
• ‘Mature’ supervisory agencies are associated
with stronger banking systems
• Potential problem
– Potential endogeneity of legal independence
– Sample size
17
Conclusions
• Reforms aimed at increasing legal independence
from government bring measurable benefits
• The chosen dependent variables are one of the
many possible dependent variables
• It is understood though that formal independence,
as important as the concept is, does not convey
the full richness of supervisory agencies’
behaviour in practice
– Detailed comparative case studies may add
value in this context
18